Abstract: Abstract: Water resources in arid region of Northwest China depend heavily on mountain precipitation and snowmelt runoff. The precipitation falling as snow or rain has different hydrological effects. As a meteorological indicator reflecting snowfall/precipitation conversion, the ratio of snowfall to precipitation (S/P) is quite important, because even small changes in S/P may influence the runoff processes and available water resources. However, the long-term trend of S/P in Tianshan Mountains, China, which is critical to the future water management in the context of climate change, has been unclear. Based on the daily meteorological data from 49 stations during 1961-2018, this study conducted a parameterization scheme to desociate the snowfall from precipitation and analyze historical spatial-temporal variations of snowfall, precipitation and S/P during cold season (from October to April) within the above timeframe. Meanwhile, considering of the impacts of climate change on S/P, five GCM models of the fifth phase of the Coupled Model Intercomparision Project (CMIP5) at RCP4.5 emission scenarios were selected to project the future change of snowfall, precipitation and S/P by bias-correction with Quantile Delta Mapping (QDM). The results demonstrated that the S/P in Tianshan Mountains was affected by the topography. The S/P was larger in the mountain area than that in the basin, greater on the northern slopes than that on the southern slopes, which was significantly and positively correlated with the altitude. During the past 58 years, the snowfall and precipitation increased significantly, and the distribution of regions with the obvious changes were concentrated in the central region of the northern Tianshan Mountains and Yili River valley. Due to the synchronous trend of snowfall and rainfall, the S/P in Tianshan Mountains was found to fluctuate between 0.35 and 0.67, revealing a slight decrease trend at a rate of 0.016%/10a. A negative correlation was also noticeable between the average wet day air temperature and S/P, and air temperature change is an important factor for S/P change. Under the RCP4.5 climate scenario, it is estimated that the snowfall in cold season will increase slowly before 2040 s±5, and then gradually decrease, showing a slow declining trend overall. In contrast, precipitation will increase significantly in the future, reaching its peak around 2080s ± 5. In this case, S/P presents a continuous declining trend during the 21stcentury, which is expected to decline from 0.52 in the 2010s to 0.43 in the 2090s. In comparison with the base period (1986-2005), snowfall in 70% sites in the Tianshan Mountains will reduced by the 2050s, with the average regional snowfall in cold season decrease by 8.9%; Whereas the precipitation of 80% stations increased, with the average regional precipitation increased by 10.1%. Affected by the amount of snowfall and precipitation, S/P of 47 out of 49 stations will decrease, and the average S/P in the 2050s will decrease by 14.7% comparing to the base period. Despite of the uncertainty in the projection results of the model, the quantitative analysis of future S/P in this study is helpful in understanding the changing process of different forms of precipitation and is of great significance to the regulation of regional water resources in relation to global warming.