Abstract: Objective The aim is to study the automated path planning method and optimization mechanism for robot weld seam inspection. Methods Through a combination of theoretical analysis and simulation experiments, based on RRT* algorithm, it explores impact of target oriented artificial potential field smoothing on smoothness and convergence efficiency of RRT* path planning optimization performance. Through mechanism derivation, simulation experiments and method comparison, probability values and expansion direction of artificial field optimization random trees are introduced. The path is modified according to robot kinematic model, and redundant points are removed by vertical distance limit method. Finally, the path is smoothed by five times B-spline method. Results The results show that compared with Goal-Biased RRT, RRT*, and APF-RRT*, SPFG-RRT* algorithm can make the path satisfy robot steering constraints and curvature constraints, with shorter path length, higher path planning efficiency and fewer turning points. The rapidly-exploring random tree algorithm has been optimized and improved. Conclusion Therefore, SPFG-RRT* algorithm improves speed and accuracy of path planning, and can meet path planning requirements of weld seam inspection robots.