The invention provides a weight-improved particle swarm optimization-based unmanned bicycle path planning method, which comprises the following steps of (1) carrying out environment modeling and coding according to a working environment of an unmanned bicycle; (2) building a fitness function; and (3) carrying out path planning of the unmanned bicycle on the basis of weight-improved particle swarm optimization. According to the path planning method, the bicycle can strictly run according to the planed path, an optimal path can be generated in dynamic and static environments, and the path planning method is few in to-be-adjusted parameters, simple in model and convenient to calculate.