<p>Extreme precipitation events pose significant challenges in South Thailand, leading to severe floods and other water-related disasters. The aim of this study was to predict future changes in extreme precipitation indices and non-stationary return levels of annual maximum daily precipitation across South Thailand using a multi-model ensemble mean of CMIP6 global climate models (GCMs). Twelve meteorological stations in South Thailand were investigated. The linear scaling method was selected as the bias correction technique. The non-stationary Generalized Extreme Value (GEV) model, with time-varying parameters, was chosen for the frequency analysis of return levels. Additionally, the Mann-Kendall test and Sen’s slope estimator were applied to assess the statistical significance of observed trends. By employing bias correction and statistical analyses, the study calculated eight extreme precipitation indices and evaluated return levels under SSP3-7.0, which reflects an emission scenario characterized by regional rivalry. The results indicated mixed trends in precipitation intensity and frequency, with significant increases in extreme indices such as Rx1day, R95p, and SDII across most meteorological stations. Non-stationary GEV models revealed an upward trend in maximum daily precipitation for future periods, highlighting the potential for intensified precipitation-related hazards. These findings underscore the importance of incorporating non-stationarity in hydrological planning and climate adaptation strategies, particularly for regions like South Thailand, which face complex geographical and climatic challenges.</p> Graphical Abstract <p></p>