In this paper, we address the capacitated vehicle routing problem CVRP. This problem consists to optimize the itineraries of the vehicles that must do with a minimum cost and it by the use of a fleet homogeneous while respecting the constraints of time, volume and capacity However, the techniques giving the optimal path are relatively gluttonous in time of calculates when the number of customers to discover is important. So and in order to palliate to this difficulty, some methods based on meta-heuristic. are used. So we opted for the ant colony algorithm (Ant colony optimization ACO); this approach is combined with a local research to improve the gotten results. This approach is applied to a real case; the society ZEDPHARM specialized in the distribution of the pharmaceutical products the used approach is simulated by Matlab.