In recent years, the application of unmanned vehicles for efficient emergency distribution has become more important in public health events, and the existing distribution methods are difficult to satisfy the needs of customers with different degrees of urgency due to limited capacity. The paper proposes a collaborative distribution mode based on unmanned vehicles and traction cabinets to solve these problems. Firstly, introducing the traction cabinet to the background of public health events, the paper proposes an emergency distribution mode of unmanned vehicle collaborative traction cabinet classified according to customer emergency degree. Secondly, the paper proposes a two-stage model combining the drop-off network location problem and routing problem, solved by an improved NSGA-II algorithm and a genetic simulated annealing algorithm. Finally, based on the relevant real data of Chaoyang District, Beijing, China, the paper designs simulation experiments. The distribution scheme solved in this paper is compared with the ordinary scheme from the aspects of path length, vehicles, and violation of constraints. It can be concluded that the model designed in the paper can better satisfy the needs of emergency distribution and achieve efficient hierarchical distribution.