Abstract: Due to the uncertainty in accident locations and demands for rescue, it is important and meaningful to locate an emergency supply stockpile warehouse. With the uncertainty in accident locations and demands for rescue considered, a robust optimization model is presented by setting a series of demanding scenarios. By this model, the distance between the warehouse and an accident location is within a given range. Then, a genetic algorithm is designed to solve the problem. It is compared with the stochastic optimization model. Result shows that the proposed method outperforms the stochastic optimization solution method. Moreover, the proposed method can effectively reduce the effect of uncertainty in the site selection.