Abstract: In order to improve the accuracy and speed of LiDAR point cloud registration, a point cloud coarse entropy that used particle swarm algorithm based on the improved beetle algorithm was adopted. The method useed the point cloud distribution entropy as the optimization target to find the optimal spatial transformation matrix, and provided a good initial condition for the precise registration of the point cloud. Through theoretical analysis and simulation verification, it was proved that the point cloud entropy is an evaluation method which has faster calculation speed than the traditional mean squared difference, and the particle swarm optimization algorithm improved by beetle antennae algorithm has the characteristics of strong global search ability and fast calculation speed. Compared with the traditional point cloud coarse registration method, the registration method has improved the registration speed by nearly 25%. Under the condition of large amount of point cloud data, it still shows a faster registration speed. This method has a reference significance for how to improve the registration speed of laser point cloud.