The present invention illustrates an automatic focusing method. Firstly, through multiple cameras of an image capturing device, a scene is captured, such that multiple images generated corresponding to the cameras are obtained. Then, multiple depth maps are generated according to the images. Next, according to the resolutions of single one or multiple objects in the depth maps, the depth information of the single one or multiple objects in the depth maps can be selected to generate a merged depth map. Then, a target focus distance of the single one object or target focus distances of multiple objects are calculated according to the merged depth map. Next, an actual focus distance of the multi-lenses module associated with the cameras is adjusted according to the target focus distance of the single one object or one of the multiple objects.