The statistical spatial‐temporal features of the intense snowfall event (ISE) in China are investigated over the period of 1962–2000. The results indicate that eastern China, northern Xinjiang, the eastern Tibetan plateau, and northeastern China are four key regions for the ISE, with more frequency and strong variability. Annual cycle analysis shows the ISE exhibits a unimodal distribution with maximum frequency at winter months for eastern China, a bimodal distribution with maximum frequency at early winter and spring months for northern Xinjiang and northeastern China, and a bimodal distribution with maximum frequency at autumn and spring months for the eastern Tibetan plateau. Linear trend analysis indicates that in the last 39 years, the ISE exhibits a decreasing trend for eastern China and an increasing trend for northern Xinjiang and the eastern Tibetan plateau. The linear trend of the ISE is weak over northeastern China. Based on the simulations of the most recent and comprehensive climate models in the 20th century run, the performance of the current climate models in simulating the Chinese ISE is investigated. The results indicate that, of the 20 models, there are four models that can reasonably reproduce the spatial‐temporal features of the Chinese ISE. Based on these four models’ simulation for the 21st century under A1B and A2 scenarios, the future variability of the Chinese ISE is projected. It is found that global warming will cause the ISE frequency over southern China to decrease, while the ISE over northern China will initially increase and then decrease.