Abstract: An open-closed mixed two-echelon vehicle routing problem (OCM-2EVRP) mathematical model is constructed based on the close-open mixed two-echelon distribution of freight in city logistics. In OCM-2EVRP, freight from a remote depot is compulsorily delivered through peripheral intermediate depots to the final customers in the city. Vehicles of the first echelon do not need to return to the depot after finishing their distribution, or they should make the same trips in the reverse order. This NP-hard problem is solved by an improved variable neighborhood search algorithm. Results of computational experiments on several extended benchmark instances show that the improved algorithm can obtain a balance between solution quality and time and outperforms an improved Memetic algorithm.