The Paris Agreement aims to keep global warming to well below 2 °C above pre-industrial levels and to pursue efforts to limit it to 1.5 °C, recognizing this will reduce the risks of natural disasters significantly. As changes in the risks of temperature extremes are often associated with changes in the temperature probability distribution, further analysis is still needed to improve understanding of the warm extremes over China. In this study, changes in the occurrence probability of temperature extremes and statistic characteristics of the temperature distribution are investigated using the fifth phase of the Coupled Model Intercomparison Project (CMIP5) multimodel simulations from 1861 to 2100. The risks of the once-in-100-year TXx and TNx events are projected to increase by 14.4 and 31.4 times at 1.5 °C warming. Even, the corresponding risks under 2 °C global warming are 23.3 and 50.6, implying that the once-in-100-year TXx and TNx events are expected to occur about every 5 and 2 years over China, respectively. The Tibetan Plateau, Northwest China and south of the Yangtze River are in greater risks suffering hot extremes (both day and night extremes). Changes in the occurrence probability of warm extremes are generally well explained by the combination of the shifts in location and scale parameters in areas with grown variability, i.e., the Tibetan Plateau for TXx, south of the Yangtze River for both TXx and TNx. The location (scale) parameter leading the risks of once-in-20-year TXx to increase by more than 5 (0.25) and 3 (0.75) times under 2 °C warming in the Tibetan Plateau and south of the Yangtze River, respectively. The location parameter is more important for regions with decreased variability e.g., the Tibetan Plateau for TNx, Northwest China for both TXx and TNx, with risks increase by more than 3, 6 and 4 times due to changes in location.