PT. XYZ is one of the manufacturing companies engaged in the transportation of trucks and buses. The company has a production system make to order with the principle of just in time and has a lead time of two months both when the order is little or much. That way, sales demand forecasting is needed to assist companies in planning the production process in the coming period. The purpose of this study is to forecast the demand for sales of Medium Duty trucks at PT. XYZ. The method used is Single Exponential Smoothing and Linear regression which later the two methods will be compared. The constants used are 0.1 to 0.9 and the accuracy measures used are MAD and MAPE. The results obtained are that the Single Exponential Smoothing method with a constant of 0.9 has the smallest accuracy size compared to others. So that the forecasting data to be used is forecasting at SES constant 0.9.