In order to solve the problem of complex and low independence of traditional tabu search algorithm,we put forward the problem of vehicle routing with time window and heterogeneous fleet on the basis of considering the vehicle routing problem with time window. In order to better solve the problem of vehicle routing with time window and heterogeneous fleet,we established a model of vehicle routing problem with time window and heterogeneous fleet considering the multiple attributes of time window, heterogeneous fleet and vehicle number limit at same time. We also proposed an improved tabu search algorithm to solve this problem. Based on the original tabu search algorithm,the reservation table,the grade cost structure principle and the vehicle sorting criterion are added in the improved tabu search algorithm to innovate it in essence. By adding the reservation table into the original algorithm and using the principle of the grade cost structure,we proposed a new algorithm to solve the vehicle routing problem. This improved tabu search algorithm solveed the shortcomings of the traditional tabu search algorithm, it not only can make the user point in the path closely arranged, but also can optimize transport routes. Finally, to calculate the effectiveness of the improved tabu search algorithm, we used the specific case data to test the calculation result using the improved tabu search algorithm. The result of the calculation proves that this innovative algorithm is effective for solving the problem of vehicle routing with time window and heterogeneous fleet.