The invention provides an unmanned bicycle-oriented artificial potential field path planning method which comprises the steps of (1) initializing environment information, determining an initial position of an unmanned bicycle, destination point position and obstacle information; (2) obtaining information of an surrounding obstacle and calculating the influence distance of the obstacle on the unmanned bicycle and the distance between the obstacle and a destination position; (3) calculating attractive force and repulsive force of the unmanned bicycle at the current position, improving a repulsive force potential field function and a repulsive force function and then calculating the magnitude and the direction of resultant force; (4) guiding the unmanned bicycle to enter the next place; (5) judging whether the unmanned bicycle falls into a local minimum point or not, if so, going to the step (6), or else, going to the step (7); (6) adding a virtual obstacle, wherein a repulsive force field is Ja, going to the step (2); and (7) if the virtual obstacle exists, deleting the virtual obstacle, judging whether the unmanned bicycle reaches the destination position or walks through the specified distance or not, if so, completing the process, or else, going to the step (2).