A method of capturing an image in a device and the device thereof are provided. The method comprises displaying a preview of an image to be captured on a screen of the device; allowing a user to execute an action input through a spatial relationship with a region, wherein the region is any area on the surface of the screen; sensing the action input; comparing the action input to a template stored in a memory of the device, wherein the template is associated with image capturing; and capturing the image if the action input substantially matches the template. The method of capturing images in the device and the device thereof provided by the present disclosure permit a user to easily capture images without clicking a fixed area on the screen.