Based on the daily 5cm soil temperature data observed by 97 meteorological stations in Shaanxi Province in 2016, combined with statistical parameters such as correlation coefficient, average deviation and root mean square error, the applicability of CMA (China Meteorological Administration) Land Data Assimilation System (CLDAS 2.0) and American Global Land surface Data Assimilation System (NoahGLDAS 2.1, NoahGLDAS1, CLMGLDAS 1) soil temperature data in Shaanxi Province was evaluated. The results show that: (1) CLDAS 2.0 had the highest correlation coefficient and the smallest rootmeansquare error in Shaanxi Province, followed by NoahGLDAS 2.1 and NoahGLDAS 1. (2) From the analysis of the time evolution series of three regions in Shaanxi Province, it can be seen that CLDAS 2.0 and NoahGLDAS 2.1 can well simulate the seasonal and daily changes of soil temperature, and the simulations of daily changes of NoahGLDAS 1 and CLMGLDAS 1 are poor, and the deviations of the former two are significantly less than those of the latter two. (3) The soil temperature simulation ability of NoahGLDAS 2.1 in the northern Shaanxi and Guanzhong area is similar to that of CLDAS 2.0, but that of CLDAS 2.0 in the southern Shaanxi area is better than that of NoahGLDAS 2.1. Generally speaking, CLDAS 2.0 has the best ability to simulate soil temperature, and has better applicability in Shaanxi Province.