Accomodation is the main area of tourism industry. In order to determine man, machine and material reguirement, right material must be on the right place on the right time. Man and material requirements are depend on room demand. Main aim of this study is to forecast hotel room demand for a five stars and international chain hotel that has been established in Ankara. In order to forecast the room demand, ARIMA, one of the most popular advanced Box-Jenkins Model, has been used. The reason of selecting this model is the consideration of time series with probabilistic and stochastic characteristics by ARIMA and performance of it for forecasts in short time periods. While analyzing, it is seen that the serie has no trend and seasonal effect but has a autocorelation. After analyses on first order differentiated data ARIMA (2,1,1) model has been found.  Test statistics for models with %95 confidence interval are; AR(1) for  p=0, AR(2) for p=0.01 ve MA (1) for p=0. The validity of the model, ARIMA (2,1,1) has been evaluated by Akaike Information Criterion (AIC). StatGraphics 16.2 and Eviews 8 has been used for analyses. After determining model, model parameters has been coded in C #. The coded model has been used as an input in a decision support system which has been used for material reguirement planning for a hotel and stored in a model base. Main role of this model is to determine master production scheduling for hotel room demand forecast in decision support system.