diff --git a/hrp/pkg/uixt/android_device.go b/hrp/pkg/uixt/android_device.go
index e2a30d93..e86206c6 100644
--- a/hrp/pkg/uixt/android_device.go
+++ b/hrp/pkg/uixt/android_device.go
@@ -310,6 +310,9 @@ func ConvertPoints(data string) (eps []ExportPoint) {
 	for _, line := range lines {
 		if strings.Contains(line, "ext") {
 			idx := strings.Index(line, "{")
+			if idx == -1 {
+				continue
+			}
 			line = line[idx:]
 			p := ExportPoint{}
 			err := json.Unmarshal([]byte(line), &p)
diff --git a/hrp/pkg/uixt/android_driver.go b/hrp/pkg/uixt/android_driver.go
index 255e9757..5ebf69dd 100644
--- a/hrp/pkg/uixt/android_driver.go
+++ b/hrp/pkg/uixt/android_driver.go
@@ -18,6 +18,17 @@ import (
 	"github.com/httprunner/httprunner/v4/hrp/pkg/gadb"
 )
 
+// See https://developer.android.com/reference/android/view/KeyEvent
+const (
+	KEYCODE_BACK      string = "4"
+	KEYCODE_CAMERA    string = "27"
+	KEYCODE_ALT_LEFT  string = "57"
+	KEYCODE_ALT_RIGHT string = "58"
+	KEYCODE_MENU      string = "82"
+	KEYCODE_BREAK     string = "121"
+	KEYCODE_ALL_APPS  string = "284"
+)
+
 var errDriverNotImplemented = errors.New("driver method not implemented")
 
 type uiaDriver struct {
@@ -233,9 +244,12 @@ func (ud *uiaDriver) Scale() (scale float64, err error) {
 }
 
 // PressBack simulates a short press on the BACK button.
-func (ud *uiaDriver) PressBack() (err error) {
+func (ud *uiaDriver) PressBack(options ...DataOption) (err error) {
 	// register(postHandler, new PressBack("/wd/hub/session/:sessionId/back"))
 	_, err = ud.httpPOST(nil, "/session", ud.sessionId, "back")
+	if err != nil {
+		_, err = ud.adbDevice.RunShellCommand("input", "keyevent", KEYCODE_BACK)
+	}
 	return
 }
 
@@ -257,7 +271,7 @@ func (ud *uiaDriver) StartCamera() (err error) {
 			return err
 		}
 		time.Sleep(5 * time.Second)
-		if _, err = ud.adbDevice.RunShellCommand("input", "keyevent", "27"); err != nil {
+		if _, err = ud.adbDevice.RunShellCommand("input", "keyevent", KEYCODE_CAMERA); err != nil {
 			return err
 		}
 		return
@@ -266,7 +280,7 @@ func (ud *uiaDriver) StartCamera() (err error) {
 			return err
 		}
 		time.Sleep(5 * time.Second)
-		if _, err = ud.adbDevice.RunShellCommand("input", "keyevent", "27"); err != nil {
+		if _, err = ud.adbDevice.RunShellCommand("input", "keyevent", KEYCODE_CAMERA); err != nil {
 			return err
 		}
 		return
@@ -578,7 +592,6 @@ func (ud *uiaDriver) _swipe(startX, startY, endX, endY interface{}, options ...D
 // per step. So for a 100 steps, the swipe will take about 1/2 second to complete.
 //  `steps` is the number of move steps sent to the system
 func (ud *uiaDriver) Swipe(fromX, fromY, toX, toY int, options ...DataOption) error {
-	options = append(options, WithDataSteps(12))
 	return ud.SwipeFloat(float64(fromX), float64(fromY), float64(toX), float64(toY), options...)
 }
 
diff --git a/hrp/pkg/uixt/ext.go b/hrp/pkg/uixt/ext.go
index 82e11553..531f95b8 100644
--- a/hrp/pkg/uixt/ext.go
+++ b/hrp/pkg/uixt/ext.go
@@ -53,6 +53,7 @@ const (
 	ACTION_DoubleTap   MobileMethod = "double_tap"
 	ACTION_Swipe       MobileMethod = "swipe"
 	ACTION_Input       MobileMethod = "input"
+	ACTION_Back        MobileMethod = "back"
 
 	// custom actions
 	ACTION_SwipeToTapApp   MobileMethod = "swipe_to_tap_app"   // swipe left & right to find app and tap
@@ -67,6 +68,8 @@ type MobileAction struct {
 	Identifier          string      `json:"identifier,omitempty" yaml:"identifier,omitempty"`                     // used to identify the action in log
 	MaxRetryTimes       int         `json:"max_retry_times,omitempty" yaml:"max_retry_times,omitempty"`           // max retry times
 	WaitTime            float64     `json:"wait_time,omitempty" yaml:"wait_time,omitempty"`                       // wait time between swipe and ocr, unit: second
+	Duration            float64     `json:"duration,omitempty" yaml:"duration,omitempty"`                         // used to set duration of ios swipe action
+	Steps               int         `json:"steps,omitempty" yaml:"steps,omitempty"`                               // used to set steps of android swipe action
 	Direction           interface{} `json:"direction,omitempty" yaml:"direction,omitempty"`                       // used by swipe to tap text or app
 	Scope               []float64   `json:"scope,omitempty" yaml:"scope,omitempty"`                               // used by ocr to get text position in the scope
 	Offset              []int       `json:"offset,omitempty" yaml:"offset,omitempty"`                             // used to tap offset of point
@@ -98,6 +101,18 @@ func WithWaitTime(sec float64) ActionOption {
 	}
 }
 
+func WithDuration(duration float64) ActionOption {
+	return func(o *MobileAction) {
+		o.Duration = duration
+	}
+}
+
+func WithSteps(steps int) ActionOption {
+	return func(o *MobileAction) {
+		o.Steps = steps
+	}
+}
+
 // WithDirection inputs direction (up, down, left, right)
 func WithDirection(direction string) ActionOption {
 	return func(o *MobileAction) {
@@ -584,6 +599,11 @@ func (dExt *DriverExt) DoAction(action MobileAction) error {
 		return fmt.Errorf("invalid %s params: %v", ACTION_DoubleTap, action.Params)
 	case ACTION_Swipe:
 		identifierOption := WithDataIdentifier(action.Identifier)
+		durationOption := WithDataPressDuration(action.Duration)
+		if action.Steps == 0 {
+			action.Steps = 10
+		}
+		stepsOption := WithDataSteps(action.Steps)
 		if positions, ok := action.Params.([]interface{}); ok {
 			// relative fromX, fromY, toX, toY of window size: [0.5, 0.9, 0.5, 0.1]
 			if len(positions) != 4 {
@@ -593,10 +613,10 @@ func (dExt *DriverExt) DoAction(action MobileAction) error {
 			fromY, _ := positions[1].(float64)
 			toX, _ := positions[2].(float64)
 			toY, _ := positions[3].(float64)
-			return dExt.SwipeRelative(fromX, fromY, toX, toY, identifierOption)
+			return dExt.SwipeRelative(fromX, fromY, toX, toY, identifierOption, durationOption, stepsOption)
 		}
 		if direction, ok := action.Params.(string); ok {
-			return dExt.SwipeTo(direction, identifierOption)
+			return dExt.SwipeTo(direction, identifierOption, durationOption, stepsOption)
 		}
 		return fmt.Errorf("invalid %s params: %v", ACTION_Swipe, action.Params)
 	case ACTION_Input:
@@ -618,6 +638,8 @@ func (dExt *DriverExt) DoAction(action MobileAction) error {
 			options = append(options, WithDataIdentifier(action.Identifier))
 		}
 		return dExt.Driver.Input(param, options...)
+	case ACTION_Back:
+		return dExt.Driver.PressBack()
 	case CtlSleep:
 		if param, ok := action.Params.(json.Number); ok {
 			seconds, _ := param.Float64()
diff --git a/hrp/pkg/uixt/interface.go b/hrp/pkg/uixt/interface.go
index c31612c6..77ae84cb 100644
--- a/hrp/pkg/uixt/interface.go
+++ b/hrp/pkg/uixt/interface.go
@@ -891,7 +891,7 @@ func NewData(data map[string]interface{}, options ...DataOption) *DataOptions {
 	}
 
 	if _, ok := dataOptions.Data["duration"]; !ok {
-		dataOptions.Data["duration"] = 1.0 // default duration
+		dataOptions.Data["duration"] = 0 // default duration
 	}
 
 	if _, ok := dataOptions.Data["frequency"]; !ok {
@@ -1046,6 +1046,9 @@ type WebDriver interface {
 	// PressButton Presses the corresponding hardware button on the device
 	PressButton(devBtn DeviceButton) error
 
+	// PressBack Presses the back button
+	PressBack(options ...DataOption) error
+
 	// IOHIDEvent Emulated triggering of the given low-level IOHID device event.
 	//  duration: The event duration in float seconds (XCTest uses 0.005 for a single press event)
 	IOHIDEvent(pageID EventPageID, usageID EventUsageID, duration ...float64) error
diff --git a/hrp/pkg/uixt/ios_driver.go b/hrp/pkg/uixt/ios_driver.go
index 4796d577..a14c08c8 100644
--- a/hrp/pkg/uixt/ios_driver.go
+++ b/hrp/pkg/uixt/ios_driver.go
@@ -440,12 +440,10 @@ func (wd *wdaDriver) DragFloat(fromX, fromY, toX, toY float64, options ...DataOp
 }
 
 func (wd *wdaDriver) Swipe(fromX, fromY, toX, toY int, options ...DataOption) error {
-	options = append(options, WithDataPressDuration(0))
 	return wd.SwipeFloat(float64(fromX), float64(fromY), float64(toX), float64(toY), options...)
 }
 
 func (wd *wdaDriver) SwipeFloat(fromX, fromY, toX, toY float64, options ...DataOption) error {
-	options = append(options, WithDataPressDuration(0))
 	return wd.DragFloat(fromX, fromY, toX, toY, options...)
 }
 
@@ -528,6 +526,27 @@ func (wd *wdaDriver) KeyboardDismiss(keyNames ...string) (err error) {
 	return
 }
 
+// PressBack simulates a short press on the BACK button.
+func (wd *wdaDriver) PressBack(options ...DataOption) (err error) {
+	windowSize, err := wd.WindowSize()
+	if err != nil {
+		return
+	}
+
+	data := map[string]interface{}{
+		"fromX": float64(windowSize.Width) * 0,
+		"fromY": float64(windowSize.Height) * 0.5,
+		"toX":   float64(windowSize.Width) * 0.6,
+		"toY":   float64(windowSize.Height) * 0.5,
+	}
+
+	// new data options in post data for extra WDA configurations
+	d := NewData(data, options...)
+
+	_, err = wd.httpPOST(d.Data, "/session", wd.sessionId, "/wda/dragfromtoforduration")
+	return
+}
+
 func (wd *wdaDriver) PressButton(devBtn DeviceButton) (err error) {
 	// [[FBRoute POST:@"/wda/pressButton"] respondWithTarget:self action:@selector(handlePressButtonCommand:)]
 	data := map[string]interface{}{"name": devBtn}
diff --git a/hrp/pkg/uixt/ocr_vedem.go b/hrp/pkg/uixt/ocr_vedem.go
index 3f456045..1bcc96ad 100644
--- a/hrp/pkg/uixt/ocr_vedem.go
+++ b/hrp/pkg/uixt/ocr_vedem.go
@@ -179,16 +179,16 @@ func (s *veDEMOCRService) FindText(text string, imageBuf []byte, options ...Data
 			}
 
 			rects = append(rects, rect)
-		}
 
-		// contains text while not match exactly
-		if ocrResult.Text != text {
-			continue
-		}
+			// contains text while not match exactly
+			if ocrResult.Text != text {
+				continue
+			}
 
-		// match exactly, and not specify index, return the first one
-		if data.Index == 0 {
-			return rect, nil
+			// match exactly, and not specify index, return the first one
+			if data.Index == 0 {
+				return rect, nil
+			}
 		}
 	}
 
diff --git a/hrp/step.go b/hrp/step.go
index 6c41f9d5..0df0347a 100644
--- a/hrp/step.go
+++ b/hrp/step.go
@@ -22,16 +22,18 @@ const (
 var (
 	WithIdentifier          = uixt.WithIdentifier
 	WithMaxRetryTimes       = uixt.WithMaxRetryTimes
-	WithWaitTime            = uixt.WithWaitTime
-	WithIndex               = uixt.WithIndex
+	WithWaitTime            = uixt.WithWaitTime // only applicable to SwipeToTap* action
+	WithIndex               = uixt.WithIndex    // index of the target element, should start from 1, only applicable to ocr actions
 	WithTimeout             = uixt.WithTimeout
 	WithIgnoreNotFoundError = uixt.WithIgnoreNotFoundError
 	WithText                = uixt.WithText
 	WithID                  = uixt.WithID
 	WithDescription         = uixt.WithDescription
+	WithDuration            = uixt.WithDuration // only applicable to ios swipe action
+	WithSteps               = uixt.WithSteps    // only applicable to android swipe action
 	WithDirection           = uixt.WithDirection
 	WithCustomDirection     = uixt.WithCustomDirection
-	WithScope               = uixt.WithScope
+	WithScope               = uixt.WithScope // only applicable to ocr actions
 	WithOffset              = uixt.WithOffset
 )
 
diff --git a/hrp/step_mobile_ui.go b/hrp/step_mobile_ui.go
index 6f9fe7fb..4dcf2be8 100644
--- a/hrp/step_mobile_ui.go
+++ b/hrp/step_mobile_ui.go
@@ -181,6 +181,18 @@ func (s *StepMobile) DoubleTap(params string, options ...uixt.ActionOption) *Ste
 	return &StepMobile{step: s.step}
 }
 
+func (s *StepMobile) Back(options ...uixt.ActionOption) *StepMobile {
+	action := uixt.MobileAction{
+		Method: uixt.ACTION_Back,
+		Params: nil,
+	}
+	for _, option := range options {
+		option(&action)
+	}
+	s.mobileStep().Actions = append(s.mobileStep().Actions, action)
+	return &StepMobile{step: s.step}
+}
+
 func (s *StepMobile) Swipe(sx, sy, ex, ey float64, options ...uixt.ActionOption) *StepMobile {
 	action := uixt.MobileAction{
 		Method: uixt.ACTION_Swipe,