This study forecasts the number of workers needed to produce predicted production quantity in the production department of a firm and makes manpower planning from current time to future production time. We used productivity for forecasting the number of workers needed and used production cost subject to learning for finding the optimal number of economic workers. This study present a method for getting the optimal number of economic workers and manpower planning under conditions of changing productivity according to service length. Simulation model devised showed to be more economical than rule of thumb practice because it could cut down wages and it could reduce costs by enabling managment to anticipate shortages or surpluses of workers. Accordingly, firm could secure the number of workers needed before shortages of workers take place.