In the event of an emergency, due to the urgency of rescue, it was very important to distribute emergency supplies to the disaster area in a timely and reasonable manner. Therefore, the decision-making of the vehicle dispatching scheme of multiple emergency supplies distribution centers became the key link of emergency logistics. Considering the uncertain travel time of the traffic network, to minimize the maximum regret value (robust cost) of the vehicle scheduling scheme, an optimization model for emergency rescue vehicle scheduling was established. Benders decomposition algorithm was designed to solve the model. First, the model was decomposed to obtain the main model and sub-model, and then the Benders optimal cut constraint was added to the solution for continuous iteration. The result of the example shows that the research model can achieve the reasonable scheduling of vehicles in multiple emergency material distribution centers with the minimum robust cost, meet the requirements of the maximum rescue time limit of vehicles, the priority of the demand of the disaster site. The vehicle travel time obtained from this model is more reasonable. This study can provide a reference for emergency logistics decision-making plans.