The purpose of this study is a model base on the combination of close – VRP and open – VRP when a company used both the private and hired vehicles to serve customers with the variant of time window that is called semi soft time window. Moreover, main point of the model is that duration of waiting time due to early arrival to customer site has a limitation. Namely, in mentioned model, is defined a pre-anticipated time windows to duration of waiting time associated with each customer. First, we have presented COMVRP problem with semi soft time window and constraints related to the passed distance for each kind of vehicle, then waiting time has been given. Then, we considered the above problem without constraint of the passing distance and the results shown reduction in number of routs and also number of active vehicles required in the problem. To this purpose, a rigorous MILP mathematical model for the COMVRP problem taking into considering Semi Soft Time Window is introduced. The objective is to minimize the total transportation costs, or similarly to determine the optimal fleet composition and dimension following least cost vehicle routes. To find solutions to this problem, we have utilized two metaheuristic algorithms, that consist of genetic algorithm and tabu search. Finally, our results show that genetic algorithm is performed well on the COMVRPSSTW