Abstract: For the innate limitations of the principle of potential field,such as no path existing among the close-spaced obstacles,oscillations in narrow passages,oscillations in the presence of obstacles,trap situations due to local minima,unreachable goals near the obstacles,etc.,this paper proposes an improved algorithm,which is adaptable to path planning of robots in the unknown complex environment and has a certain degree of learning ability. The effectiveness of this method is verified by simulation results.