A method for processing an image in a virtual reality display device comprises determining a gaze area and a non-gaze area in a display area of the virtual reality display device, performing, at a first time instant, a rendering process in the gaze area and the non-gaze area to generate a first image, performing a time warping process on the first image to generate a second image, based on attitude information of the virtual reality display device at a second time instant, and modifying the second image based on the movement information and attribute information of a motion object in the gaze area. The method can avoid problems caused by the time warping process, reduce delays, and avoid smearing or ghosting, and thus obtain a good display effect.