The invention discloses a dynamic path planning method for an improved particle swarm algorithm, which is characterized by comprising the following steps of: 1, generating a global path by adopting the improved particle swarm algorithm, and dividing the global path into n sections of local paths consisting of {S1G1, S2G2,... SnGn}; step 2, performing path planning on the local path S1G1 by using a dynamic window method DWA, including building a robot kinematics model, building a robot speed model, building an objective function of the dynamic window method DWA, selecting an optimal track according to the objective function, and performing path planning on the local path S1G1 by using the optimal track; step 3, performing local path planning on the local path S2G2 by using a dynamic window method DWA, and inheriting the course angle of the previous motion planning; 4, the dynamic window method DWA is repeatedly used for conducting local path planning on the paths S3G3,..., SnGn in sequence; and step 5, outputting a complete final path. The method has the characteristics of shortening the length of the planned path, improving the smoothness and real-time performance of the planned path and the like.