To address the issues of sampling point redundancy,low-quality path planning,and excessive path curvature in the Bidirectional Rapidly-exploring Random Tree (Bi-RRT^*) algorithm for mobile robot path planning,an improved APF-Bi-RRT^* algorithm is proposed.First,a dynamic target bias function is introduced into the Bi-RRT^* to reduce the number of sampling points and improve the path planning efficiency.Second,a dynamic repulsion coefficient is incorporated into the Artificial Potential Field (APF) to enable real-time adjustment of the repulsive force,thereby enhancing the quality of the path planning.Finally,the pruning algorithm is improved by integrating a safe distance detection strategy combined with cubic B-spline curves to alleviate the path curvature.The proposed algorithm was validated through simulation experiments in diverse environments.The results show that it improves planning efficiency by 16.46%,22.52%,and 62.68% compared to the APF-Bi-RRT^*,Bi-RRT^*,and RRT^* algorithms,respectively.The improved algorithm significantly reduces computation time and the number of sampling points,while achieving shorter path lengths and higher smoothness.