In this paper, computer image processing and artificial neural networks were used in detecting and grading system of agricultural products such as apple, peach and eggs. It can detect and grade agricultural products accurately according to color, shape, size and defect by image acquisition and analysis system and corresponding feature database. Using the system, the accuracy rate of grading 100 Fushi Apples according to their quality reached 96 %. By developing corresponding feature database, the system can be applied into other agricultural products.