When natural disasters occur, it is of great importance to distribute life relief supplies to disaster relief sites as soon as possible. In order to optimize the emergency material distribution process under natural disaster conditions, truck-Unmanned Aerial Vehicle(UAV) was used for common distribution, which transfered the emergency material distribution problem under that condition to a multi-traveler problem with unmanned aerial vehicle. Considering the accessibility of the truck, a mixed integer programming model was established with target of the shortest total delivery time, and the FSTST (Flying Sidekick Traveling Salesman Problem) algorithm was improved to solve the model. The simulation result shows that considering the accessibility of truck, the problem of emergency material distribution with multi-truck and multi-unmanned aerial vehicle can be effectively solved through the combination of K-means clustering algorithm and improved FSTSP algorithm. And compared with traditional truck distribution, truck-UAV distribution has potential advantages.