Demand planning is an integral part of any planning process. Accurate forecasts help firms effectively plan the production process so that inventory levels in the supply chain can be optimized and supply can be matched closely with demand. Demand planning can also help the marketing department of a firm to decide upon the kind of promotional exercises required for a particular product. Planning accurately leads to better distribution planning as well. Since firms can determine the exact levels of inventory to be held at each distribution center. In this paper, an attempt was made to forecast the demand of Automotive Batteries. Three different methods of forecast have been used. After applying those methods, finally the mean square error was minimized and the optimum weights to the forecasts was assigned by the different methods and the resultant forecast combining all the forecasts was found out. A suitable tool for the optimization was chosen. Here, Genetic Algorithms have been chosen for obtaining optimal weights that are assigned to forecast methods to generate a model of the forecast with minimum mean square error. An extensive computational experience has been reported. The proposed methodology has been put into use in the firm for better forecast of the demand.