Logistics distribution has become the key research aspect to improve the efficiency and reduce the cost of logistics. Based on the survey of current situation and optimization algorithms, a novel optimization scheme is presented in this paper. For a number of distribution sites in a city, firstly K-means clustering algorithm is adopted to examine local distribution centers and their scope. Following this, ant colony algorithm is also used to design the local optimal route inside each scope. The simulation results show that the presented scheme can improve the optimal distribution route compared to pure ant colony algorithm.