By means of the observation data from the Aletai station of the northern Xinjiang from 1954 to 2010, the probabilistic risk analysis for snow disasters is conducted for five major climate factors, and based on the fuzzy information diffusion theory, a comprehensive analysis experiment is carried out. The results show that in the case of unknown theoretical distribution, for the probability risk analysis of snow climate factors, the fuzzy information diffusion method is simple and reliable, but rough. The theoretical analysis of distribution indicates that the theoretical models of the number of heavy snow days in winter (≥6 mm), maximum snow depth, maximum daily snowfall, snowfall in winter, and the number of snow days with snowfall greater than or equal to 10 cm have the distribution characteristics of Gamma, with a confidence degree of 0001, passing through the test of significant correlation coefficient. The probability density distributions of climate factors are analyzed for the numbers of the days with different boundary snow depths. The results exhibit the two peak characteristic, which does not belong to the conventional canonical distribution, and the theoretical distribution needs further study. The proposed snow climate factors with gamma distribution are of significance for understanding the climate risk of snow disasters in theory.