Mountain snow is a fundamental freshwater supply in the arid regions. Climate warming alters the timing of snowmelt and shortens the snow cover duration, which profoundly influences the regional climate and water management. However, a reliable estimation of snow mass in the Tianshan Mountains (TS) is still unclear due to the scarcity of extensive continuous surface observations and a complex spatial heterogeneity. Therefore, a long-time series of snow simulation was performed in the WRF/Noah-MP from 1982 until 2018 to quantify the snow mass in the TS, forced by the ERA5 reanalysis data and real-time updated leaf area index and green vegetation fraction. Meanwhile, March snow mass (close to the annual peak snow mass), snow cover fraction (SCF), and trends were investigated in the TS. The results indicated a good accuracy of the estimated snow water equivalent (root mean square error (RMSE): 7.82 mm/day) with a slight overestimation (2.84 mm/day). Compared with the ERA5 dataset, the RMSE and mean bias (MB) of the daily snow depth from the WRF/Noah-MP were significantly reduced by 95.74% and 93.02%, respectively. The climatological March snow mass measured 97.85 (±16.60) gigatonnes in the TS and exhibited a negligible tendency. The total precipitation during the cold season controlled the variations of the March snow mass. The increased precipitation in the high-altitude regions contributed to an extensive snow mass, which could offset the loss in the TS lowland. In contrast, rapidly rising air temperature caused a significant reduction of the March SCF, particularly in the Southern TS.