The traditional assembly system relies on manual labor for repetitive labor,which is easy to cause errors due to the fatigue and limited human eye re solution of human operations.In order to avoid wasting labor,time and solve un stable factors such as light in the factory environment,it is proposed a method for recognizing industrial parts with various shapes based on YOLO v3algorithm .In the intelligent assembly system,the types of parts are judged based on the visual inspection results,which makes up for the shortcomings of the tradition al methods and meets the beat requirements of the product production system.The improved YOLO v3network model uses the k-means algorithm to cluster the param e ters of the anchor box,and the residual network to reduce the network parameter s.Combined with the multi-scale method and the Mish activation function to imp r ove accuracy,it is more suitable for small industrial parts and target classifi cation detection.The model uses 3D printed industrial parts to make a data set. Experiments show that compared with the original YOLO v3algorithm,using the i mproved network model has good robustness,the accuracy rate is increased by 1.52%,and the time is increased by 7.25ms.Accurately detect the types of parts in the intelligent assembly system in real time.