Objective: To improve the accuracy and comprehensiveness of oyster grading. Methods: The oyster automatic grading equipment was proposed and designed, the oyster queuing structure combining the rotating drum and the baffle conveyor belt, the grading method combining weight detection and machine vision detection were determined, and the overall structure design of the oyster grading equipment was completed. The oyster image was collected by industrial camera, and the oyster image was extracted by Otsu binarization, Gaussian filtering processing, Canny operator edge extraction and other methods. The oyster was graded by machine vision algorithm with length and fullness as the standard, and the comparison test between machine vision grading and manual grading were carried out. Results: The machine vision classification accuracy of oysters was 95.4%, and the image detection speed was about 0.647 s/image. Conclusion: Machine vision is effective for oyster grading and can classify oysters more accurately.