Abstract: In order to solve the problem of low generality of traditional machine tool fixtures, errors and low efficiency caused by multiple tool settings and installation, an improvement scheme of machine tool clamping based on point cloud reverse modeling technology is proposed. By installing a laser scanner on both sides of the fixture to obtain the 3-D point cloud information of the workpiece and the fixture, the curvature discrimination algorithm and iterative closest point algorithm based on particle swarm search were used to denoise and coordinate registration, and the 3-D space coordinates of the workpiece and the machine tool were established. Finally, the offset between the space coordinate system and the machine coordinate system was calculated according to the point cloud coordinate of the reference point, and the offset value was entered into the computerized numerical control system for machining. The experimental results show that the improved fixture can effectively reduce the number of tool settings and clamping, and the overall efficiency is increased by about 30%. The research has a certain reference value for improving the processing efficiency of parts with more machining surfaces and more complicated processes.