Crop is one of the food crops that our people depend on for survival. The occurrence of diseases will cause the decline of yield and quality, so it is very important to timely and accurate control of diseases. With the development of computer vision, crop disease recognition based on image has become the development trend of crop disease recognition. Firstly, this paper briefly introduces the related theories of image recognition, and then introduces in detail the research progress of crop disease recognition methods based on traditional machine vision technology, deep learning technology, deep learning and transfer learning at home and abroad, and analyzes the advantages and disadvantages of each recognition method, It is pointed out that the recognition methods of deep learning and transfer learning will become the main development direction of crop disease research.