The article studies the classification method for the fresh agaricus bisporus based on image processing. When acquiring the image information, the shadow of image and mushroom stipe may affect the analysis of maximum diameter of agaricus bisporus which is the important factor. In this paper, the global threshold segmentation method and maximum entropy threshold segmentation method are combined to carry out the first watershed algorithm to remove the shadow of image. Then the Canny operator, opening and closing operation and corrosion expansion are used to carry out the second watershed algorithm for the removal of stipe interference. The method achieves a good result through the comparison between the actual measured results and experimental results.