Abstract: Surface temperature, as a critical parameter in land surface processes, can alter local atmospheric circulation and cause regional short-term climate anomalies by influencing soil thermal properties and moisture exchange. To evaluate the applicability of different surface temperature reanalysis products over the Tibetan Plateau, this study systematically compares the bias characteristics of surface temperature between in-situ observations and two reanalysis datasets (CRA40 and ERA5) from 1980 to 2019. The results indicate that: (1) Both CRA40 and ERA5 reanalysis datasets can reflect the distribution of surface temperature, which gradually decreases from the southeast to the northwest, as well as its interannual variation trend and abrupt changes. However, certain interannual and interdecadal discrepancies exist in the simulation of observed values in the reanalysis data. (2) Surface temperature derived from CRA40 and ERA5 generally presents cold biases with pronounced regional and seasonal heterogeneities. Specifically, larger biases are mainly located in the southeastern and southern marginal areas of the plateau, and the magnitude of cold biases are larger in January than in July. (3) Both CRA40 and ERA5 reanalysis datasets are significantly positively correlated with the observations. Overall, CRA40 performs better than ERA5 and is more suitable for climate change research over the Tibetan Plateau, whereas relevant seasonal and regional discrepancies should be fully considered in future studies.