Abstract: Hospital site selection is crucial to modern hospital construction. This paper discussed the stochastic characteristic of parameter (monthly visits) and proposed a stochastic programming based method to deal with the problem of hospital site selection. Various factors are taken into consideration, and a stochastic integer model is built with the objective to minimize total weighted distance between population and hospitals. Monte Carlo method is used to simulate stochastic parameter. The computational experiments prove the good convergence of the solution and the improvement of site selection decision compared to the deterministic model.