In order to cope with the complex and changing environment, it is important to modify and optimize the auto-driving algorithm, making it more efficient and effective. For example, how to calculate an optimum path in congested roads. Through an in-depth study of traditional A* and D* path planning algorithm, a smoothness improvement for them is indispensable. This research innovatively combined the A-star algorithm with Bezier curves and the artificial potential field method, and the combination of them was applied to the generated obstacle scene to precisely meet the design requirements. In the face of more complex pathfinding scenarios, the combination of D-star algorithm and spline interpolation was used to address the problem. The experimental results show that under this comprehensive algorithm strategy, the robot’s turning smoothness was significantly improved, and it shows excellent real-time obstacle avoidance ability. This paper carries out simulation experiments with MATLAB to fully verify the feasibility and effectiveness of the algorithm.