The classification of hemp walnuts helps in the sale of products, and the traditional classification is limited to manual operations. In order to realize the automatic classification of hemp walnut, a hemp walnut classification algorithm was designed, which was realized by constructing a walnut pixel probability distribution model. Specifically, based on different views of the walnut, a pixel probability distribution model is constructed using the same type of walnut. The walnut to be tested is compared with the model, and the ratio of the weight inside and outside the contour of the walnut to be tested is taken as the similarity between the walnut and the model, and the walnut is classified. Through the establishment of a data set of 9 000 images, which contains three types of walnuts, lion head, tiger head and official hat, the performance of the algorithm is tested. After testing, the algorithm achieved an average recognition rate of 97.36% in 3 cross-test experiments.The experimental results show that this method has a good application prospect of the classification and identification of hemp walnut.