Abstract: The traditional artificial potential field (APF) algorithm has drawbacks such as target inaccessibility, local minima, and low planning efficiency in the process of path generation. To address these issues, this paper proposes an improved APF algorithm. For the problem of target inaccessibility, a new repulsive field function is designed, which dynamically adjusts the repulsion exerted by obstacles near the target point on the robot by introducing a sine distance factor. For the local minima problem, an tangent vector algorithm of obstacle boundary point group is proposed, which constructs a virtual point group using the obstacle boundary to calculate temporary target points for escaping local minima. For the problem of planning efficiency, a method of adaptive step size is adopted, which adaptively selects the iteration step size according to the congestion degree of obstacles around the robot. This not only reduces the number of iterations, but also prevents the robot from being repelled due to being too close to obstacles. Simulation results show that the improved algorithm can effectively solve the problems of target inaccessibility and local minima in the traditional APF. In scenarios without such problems, the path length of the improved algorithm is close to that of the traditional APF, while the planning time is shortened by approximately 23.52%. Through horizontal comparison experiments with the current mainstream robot local path planning algorithms, the superiority of the proposed algorithm in terms of planning efficiency and path quality is further verified. Finally, experiments on real robots confirm the feasibility and effectiveness of the algorithm.