Abstract: An improved Particle Swarm Optimization (PSO) algorithm is designed for TSP solving based on the information communication and dynamic work allocation. A strategy of information communication among particles with greedy idea is proposed so as to make a particle gain more useful information from other particles. The general algorithm frame of PSO is also enhanced, combined with dynamic work allocation strategy to keep the balance between the searching efficiency and solution quality. Some experiments are conducted and demonstrate that our improved PSO algorithm for TSP problem is efficient and promising.