Abstract: Because the information structure of conventional demand forecasting model for supply chain is in discrete state and the prediction structures of individuals are different greatly from each other, the inventory replenishment strategy is seriously affected. So, the policy CPFR (collaborative planning, forecasting and replenishment) is introduced firstly into the study on collaborative replenishment to develop the mathematic models of the function of optimum inventory in terms of optimum delivery period. Then, an evolution strategy (ES) is used to give the solution to the models. The results indicate that total cost increases with increasing parameter values, namely, the optimum delivery period and the optimum inventory are affected with the different parameter values. The conclusion verifies the validity of the policy proposed.