Abstract: Based on a combination of historical runs, one of the core experiments of the 5th Coupled Model Inter-comparison Project (CMIP5), and remote sensing data, simulations of the snow cover fraction over Eurasia from 13 coupled climate models were evaluated, using correlation analysis, root-mean-square error (RMSE), standard deviation, and other statistical metrics. The ensemble mean of the 13 models was used to project changes in snow cover fraction (SCF) over Eurasia under the conditions of three different representative concentration pathways (rcp2.6, rcp4.5, and rcp8.5). The results show that, although large differences are seen among simulations and between simulation and observation on the Tibetan Plateau, on the whole, the models were able to adequately simulate the spatial patterns, seasonal changes, and annual variation in SCF. The projection scenarios indicated a clearly decreasing trend from 2006 to 2040 under the different emission scenarios, with little difference between them. However, after 2040, the decreasing trend tends to be smoother under rcp2.6 and rcp4.5, whereas under rcp8.5 the trend is less pronounced. In this, the areas experiencing the most significant decrease are located to the west of Europe and on the Tibetan Plateau, and the greatest decrease is seen in winter, spring, and autumn. The results suggest that to maintain Eurasian snow cover, it is important to control greenhouse gas emissions in the future.