Abstract: The average snow depth, snow day, snow disaster periodicity, slope, livestock density, overloading rate of livestock in winter and spring, grass production, regional GDP and net income of the farmers in 2010 were selected as the main risk factors of snow disaster. A regression model of integrated risk assessment of snow disaster was established and the integrated risks in the Three Rivers Source Region were systematically evaluated using logistic regression methods and the ArcGIS and SPSS software tools. 1) During 1960-1980, the frequency of snow disaster was relatively lower in the Three Rivers Source Region in winter and spring. After 1980, the frequency showed a trend to increase. During 1980-2008, the numbers of snow disasters accounted for 62% of the total numbers in the past 59 years. In spatial distribution, snow disasters were mainly concentrated in the southeastern area of the Three Rivers Source Region. 2) Regression results of integrated risk showed that the average snow depth, snow disaster periodicity, grass production and livestock density had higher importance or with regression coefficients of 2.17, 1.38, 1.27 and 0.92 respectively, while net income of the farmers had lower importance than other factors. 3) The areas with higher integrated risks of snow disaster were located mainly in Yushu, Chengduo, Zaduo and Nangqian counties in southern regions of the Bayan Har Mountains and Gande, Dari, Maqin and Jiuzhi counties between the A'nyemaqen Range and Bayan Har Mountains A'nyemaqen Range in 2010, while the areas with lower integrated risk were concentrated in most areas of no man's land in Hoh Xil and in the Togton river basin. 4) Assessment results of integrated risk of snow disaster accorded with the actual distribution of snow disasters in the Three Rivers Source Region in the past 59 years. This study is of great significance for mitigating the losses from snow disasters and ensuring the sustainable stockbreeding development in pastoral areas. It can provide a scientific basis for mitigation planning pre-disaster and for decision strategies of assistance and recovery post-disaster for the local government departments.