The invention provides a method for improving autostereoscopic display resolution ratio. The method includes that disparity images of display scene information are recorded and stored in a multi-channel video format; in the process of generating of a synthesis image, (i)n(/i) disparity images can be acquired each time by only separating and parallel decoding, and then a multi-viewpoint synthesis image can be generated. Compared with a method for generating a synthesis image through a multi-lattice video source, the method has the advantages that the step of zooming is not needed, in other words, a display value of each sub-pixel in the synthesis map can accurately reflect recorded scene information, so that display resolution ratio is increased.