This study examines some commonly used forecasting methods of tourism demand in Macau. The main purpose is to complement the insufficient empirical works in Macau tourism forecasting. Various well-known forecasting methods, including naïve, moving average, time series decomposition, exponential smoothing, autoregressive integrated moving average, time series regression, and artificial neural network, are employed in this study. Monthly tourist arrival data from January 2001 to December 2010 are used to build the models. To select the best model from each method, accuracy of the tourist forecasts of each model is evaluated by the mean squared error (MSE) using in-sample (i.e. ex-post) data. Then, the forecasting performance of the selected models is compared by the mean absolute deviation (MAD), mean squared error (MSE), mean absolute percentage error (MAPE), and the Theil’s U statistics using out-sample (i.e. ex-ante) data. The empirical results show that sophisticated models do not necessarily produce more accurate forecasts than simple models. The simple forecasting methods such as naïve, exponential smoothing, and simple time series regression could outperform the sophisticated methods such as autoregressive integrated moving average and artificial neural network. Moreover, the inclusion of the trend and seasonal components in the models generally enhances the forecasting performance due to the similar pattern of the tourist arrivals to Macau. Furthermore, this study finds that the impacts of the financial tsunami in late 2007 does affect the forecasting accuracy of the models because of the discrepancy of the results between the two different model estimation periods. All these findings should contribute to Macau economic policy planning and sustainable growth of Macau tourism industry