The work aims to enable the bat algorithm (BA) to be applied to solve the vehicle routing problem (VRP) in package distribution, and improve the solving performance of BA. Based on the standard BA, a hybrid BA (HBA) was proposed. Firstly, the improved BA (IBA) was designed to solve VRP in package distribution. Secondly, the chaotic system was introduced to initialize IBA. Then, the fission operator and mutation operator were designed. In order to improve the population diversity, the worse half bats in the group were reinitialized by chaotic system during the first half iteration. In the second half iteration, the catfish disturbance was carried out on the bats fallen into the local optimal solution. Finally, the HBA was put forward and simulation tests were carried out on an enterprise instance. The optimal distribution distance obtained by HBA was 773.01 km, which respectively saved 8.24 km and 13.03 km compared with the GA (781.25 km) and BA (786.04 km). Compared with IBA and GA, the HBA has stronger global optimization ability and faster convergence speed when used to solve the VRP in package distribution.