In this paper two popular time series methods for modeling seasonality in tourism forecasts are compared. The first uses a decomposition methodology to estimate seasonal variation. In this method seasonal variation is estimated with a ratio-to-centered moving average approach. Three different approaches in calculating the seasonal indices are analyzed. The deseasonalized series are then forecast using an ARIMA model. The second methodology uses a multiplicative seasonal ARIMA (SARIMA) approach to simultaneously model trend and seasonal variations. The two methodologies are compared and the accuracy and managerial advantages of each are discussed.