Abstract: Aiming at the problems of easy crossing obstacles and sharp path inflection points in path planning by jump point search (JPS) algorithm, a path planning algorithm based on improved JPS with cubic B-spline interpolation was proposed. Based on the original JPS algorithm, the path inflection points near obstacles were conditionally restricted to reduce the possibility of crossing obstacles. The original cubic B-spline interpolation algorithm was improved by introducing a time axis to further optimize the path inflection points and make the path smoother. Based on the simulation verification of the proposed algorithm, it was registered as a plug-in to the ROS for path planning of the robot. The results show that the improved JPS algorithm can effectively improve the robot’ s ability to avoid obstacles on the basis of retaining the original search efficiency. The introduction of the time axis can improve the path sharpness, and the optimized path is more consistent with the actual motion planning of the robot. In the robot path planning of real scenarios, the robot can successfully reach the set target point from the specified starting point, and its search time is reduced by about 20% compared with that of A* algorithm, which effectively improves the accuracy and real-time of robot path planning.