This paper presents a linear integer programming model for a capacitated backhaul vehicle routing problem with time window (CVRPBTW) that minimizes the fleet cost and total traveled or the duration of the tours. We consider a heterogeneous fixed fleet with a fixed number of vehicles for each type (class). It is shown that the presented model enables to establish routes to serve all given customers (i.e., nodes) by the minimum number of vehicles with the maximum capacity used and minimum duration. This model is solved by a hybrid simulated annealing (HSA) method, in which is able to find very good solutions in a reasonable amount of time. A number of test problems in a small and large size are solved, and the computational results are reported in this paper. We investigate the performance of the proposed method in a real case. The input data are collected from waste and garbage collection industries for specified areas in Tehran.