The vehicle routing problem with time windows (VRPTW) is based on a vehicle routing problem (VRP) which increased customer access time windows. More and more of the VRPTW research is attention because many real-life problems can be attributed to VRPTW, Consequently, good vehicle routing can not only increase the profit of logistics but also make logistics management more scientific. In view of the characteristic of logistics, an improved ant colony algorithm is proposed. The new algorithm is proposed to fast solve the initial solution to accelerate the convergence speed, and used pheromones adjust strategies to prevent falling into local optimization in order to improve search results. The path length is improved through path construct strategy and the introduction of the exchange law to reduce the computation, thereby increasing search speed. The algorithm given in this paper can obtain an optimized solution effectively and has proved to be a good scheme to solve VRPTW.