With the development of robot technology, the manipulator has been used in various fields. A safe and effective motion path is the key to the application of manipulator. So that the path planning algorithm for multi-target point traversal of manipulators is studied and implemented on a 6-DOF (degree of freedom) manipulator in this paper. Firstly, for the general working scene of manipulators, the working space is divided into obstacle-free space and operation space, and the RRT (rapidly exploring random tree) algorithm is combined with joint trajectory planning algorithm. It will effectively reduce the search space of the RRT algorithm, thereby the planning efficiency is improved. Considering the general task requirements which is moving the end of the manipulator between multiple points, this paper proposes an efficient improved RRT algorithm to realize the construction of multiple random trees at the same time and effectively connect them. It can improve the utilization of random trees and the efficiency of path planning between multiple points. On this basis, the result of path planning is shortened by reducing the waypoints. So that the algorithm result is optimized. Finally, the effectiveness and advancement of the improved algorithm for multi-point traversal path planning are verified by simulation.