Aiming at the situation of grasping the disorderly industrial parts in the production process of industrial parts,a new method of pose estimation based on weighted voting was proposed.Introducing the weighted voting method based on the original point-pair feature (PPF) algorithm could solve the misvoting problem caused by a large number of shortline point pairs on the model surface at the offline modeling stage,realize the point cloud registration of the target work pieces,and finally get the pose estimation results of the target parts.Compared with the reconstructed model with 3D scanning equipment,it had been proved that the proposed method in this study could carry out registration experiments,reduce the workload at the offline modeling stage and well adapt to the working situation of the manipulators grasping industrial parts.