Abstract：Aiming at AGV’s single path and multi-path cross path identification problem, an AGV path recognition method based on different color discrimination is proposed. In the industrial site, two different colors of red and yellow ribbons are used to lay the path, the RGB format images collected by the camera are converted into grayscale images and HSV format images, and then the grayscale images and the S channel images are mixed. After the image is threshed, the binary path image is obtained. According to the difference of red and yellow road image characteristics and navigation commands when crossing multiple paths on a single path, different road types are recognized. And then to control the AGV by calculating the direction offset and angle offset of the path centerline. Experiments show that this algorithm of path recognization is simple and robust, and the average time for processing a frame of image is much shorter. The method combines the color recognition results of the path to provide a solution to the path recognition problem when a single path and multiple paths cross. Finally, the effectiveness of the proposed method is verified by experiments.