Abstract: In order to reach a global optimization in multi-depots vehicle routing,vehicle routing models based on detail order information were established.Hybrid algorithm was composed of particle swarm optimization(PSO) and improved ant colony optimization(ACO).Order numbers for vehicles to freight were got by particle position vector,single vehicle route was got by ACO,and then evaluated and filtered particales according to optimal vehicle routes,circulated until terminate qualification.By optimizing all vehicles routing to all orders,the model and hybrid algorithm solves the problem of searching local optimal solution in the procedure so that multi-depots transform to many single depots.Illustration results showed vehicle route length by the hybrid algorithm to be less than ant colony optimization.