Abstract: The quality of grain appearance is a crucial foundation for purchasing and grading, and the precision of inspection results is directly related to the consumers and sellers. The rapid non-destructive detection is of great significance to enhance the efficiency of grain purchasing, reduce the labor intensity and improve the detection ability of grass-roots grain enterprises. Meanwhile, machine vision technology is not only increasingly mature in the research of agricultural non-destructive testing, but also gradually in-depth in application of grain appearance quality detection. This paper summarizes the achievements and applications of image acquisition, image processing, feature extraction and classification algorithms in the application of grain appearance quality detection, and discusses the development status and shortcomings of the above methods, based on the comparative study of appearance quality detection algorithms. Meanwhile, the development trend and future research direction of this field were proposed.