In this study, soil moisture data from two reanalysis datasets (ERA‐Interim, ERA5), a satellite soil moisture product from the European Space Agency (ESA) and three assimilation datasets from the Global Land Data Assimilation System (GLDAS), that is, GLADS‐NOAH025, GLADS‐NOAH10, and GLDAS‐CLM were compared with the observed soil moisture data in Maqu and Maduo stations over the Three Rivers Source Region (TRSR) on the Tibetan Plateau in China. Comparative statistical parameters between the soil moisture observations and the products, including the root mean square error, correlation coefficient (R), and standard deviation ratio were calculated to evaluate the products in different seasons. It was found that most products overestimated the soil moisture values, among which GLADS‐NOAH025 had relatively good consistency with the observations in the surface soil layer (0–10 cm), whereas GLDAS‐CLM agreed well with the observations in the 10–40 cm layer. Both ERA5 and ERA‐interim soil moisture overestimated soil moisture at all levels. The responses of soil moisture to regional climate change over the TRSR were also investigated by using the GLDAS‐NOAH025 soil moisture product, the ESA soil moisture product, and the observations during the nonfreezing period. The results show that the trends of soil moisture are consistent with the precipitation changes in the main central part of the TRSR. A significant positive correlation was found between soil moisture and precipitation in most regions of the TRSR, which indicates that soil moisture increases with the increase of precipitation, except in the northwest TRSR, where the variation of soil moisture was possibly influenced by both the precipitation changes and the increase of evapotranspiration.