Abstract: Taking the shortest total travel time as the optimization object, a model of single objective programming is constructed, which contains a series of constraints, such as single depot, single vehicle type, loading capacity limitation, split deliveries by backpack and so on. An adaptive tabu search algorithm (ATSA) is designed to solve the problem. The ATSA uses a series of optimization strategies to improve the search ability, such as an adaptive penalty mechanism and a multi-neighborhood structure, and designs corresponding neighborhood operators for both customer points and backpack, which can better adapt to the discrete splitting degree of customer demands. The validity of the proposed model and algorithm is demonstrated by the comparison of the test results with the literature.