Objective To solve the problems of low search efficiency and numerous path inflection points of the traditional A* algorithm in complex environments an improved bidirectional A* algorithm is proposed. Methods This algorithm dynamically expands target points in both forward and reverse directions employs dynamic heuristic functions and improves the search area from an 8-neighborhood to 24-neighborhood in 8 directions. To optimize the path multiple thirdorder Bézier curves are introduced for path smoothing. Results In scenarios with multiple obstacles and complex maps the improved algorithm shows higher efficiency. During path planning it can quickly find the optimized path reduce the number of searched nodes and path inflection points and the experimental data index is improved by more than 85%. Conclusion The improved bidirectional A* algorithm has better adaptability and flexibility making it suitable for path planning in various complex environments. Its performance advantages make it an ideal choice for handling large-scale and highly complex maps and this algorithm provides strong support in the field of path planning.