Based on the daily average temperature and surface (0 cm) temperature data observed by 71 meteorological stations in Zhejiang Province in 2016, the applicability of ERA reanalysis data of land surface temperature is evaluated. By calculating the correlation coefficients, mean deviation, mean absolute deviation, root mean square error and Nash efficiency coefficient between the two sets of reanalysis data (ERA5 and ERAInterim) and the observation data, the ERA reanalysis data is comprehensively evaluated. Both the spatial and temporal distribution characteristics and changes of 2m temperature in Zhejiang Province can be better reproduced. The correlation coefficient is higher than 0.98, and the absolute deviation is smaller. For surface temperature, the applicability of the two sets of reanalysis data is worse than that of air temperature, mainly because in the two sets of reanalysis data, the surface temperatures are lower than the observed values, and the deviation in summer is significantly greater than in other seasons, but the correlation coefficients of most sites are higher than 0.9, while the root mean square error is higher than 2 ℃. In general, the two sets of reanalysis data of land surface temperature have good applicability in Zhejiang Province. ERA5 is better than ERAInterim as a whole and the improvement of surface temperature is more obvious.