The aim of this study is to model and forecast the tourist arrivals from East Asia, namely China, Korea, and Japan, to Thailand for the period from 1991 to 2016. In order to achieve this, two forecast models are applied: the AR(m)- GARCH(p,q), and the Kink AR-GARCH model (Kink AR(m)-GARCH(p,q)) that combine the classical GARCH model of Bollerslev (1986) with the Kink model of Chan and Tsay (1998). The accuracy of the forecast models is evaluated in terms of the RMSE, the MAE and the MSPE. The empirical results show that the Kink AR(1)-GARCH(1,1) produces forecast which perform(statistically) significantly better than AR(1)-GARCH(1,1) in forecasting tourist arrivals from China and Korea to Thailand. However, AR(1)-GARCH(1,1) is preferred for forecasting international tourism demand for Thailand from Japan. © 2016 by the Mathematical Association of Thailand. All rights reserved.