Abstract: In order to facilitate the subsequent processing, it is necessary to develop an identification system which was used to classify the freshwater fish automatically by their species. In this study, four species of fish amounted 240 for the test samples were collected, which were chub, crucian, bream and carp, the image of each fish by the machine vision were got, after the image processing, the color components and the ratio between the major axis and the minor axis of the fish were extracted. Finally the identification model was extracted by the eigenvalues. The results showed that the identification model would be used to identify the four species; the precision rate was 96.67%. Computer vision can be used to grade the freshwater fish quickly and precisely, it has great practical application value.