This paper divides the moon's appearance by studying the texture characteristics of the gray-scale image of the moon captured by Chang'e-1. This paper proposes a method of using the lunar texture features combined with the gray value of the lunar image and adopting the Bayesian classification method to classify the lunar landforms. The texture features of the lunar image are described by 13 texture features calculated by the gray-level co-occurrence matrix. The specific method is to first select the best texture features that can distinguish different lunar landforms and the corresponding optimal window size used to extract these best texture features, and then perform principal component analysis on these extracted texture features to remove the correlation, And then use Bayesian classification to classify the appearance of the moon. Experiments show that this method can well extract the texture features of the lunar surface, and can successfully automatically identify and classify lunar landform units. The finished lunar geomorphological zoning map can provide reference for the further study of the moon and provide detailed information for better exploration of the moon.