To provide guidance for the use of reanalysis data in the Central Taklimakan Desert (CTD), multiple reanalysis products are evaluated based on the radiosonde data obtained from two field experiments for the first time in the CTD, including the European Centre for Medium-Range Weather Forecasts (ECMWF) Reanalysis version 5 (ERA5), ECMWF Reanalysis-Interim (ERA-Interim), Japanese 55-year Reanalysis (JRA55), Modern-Era Retrospective Analysis for Research and Applications version 2 (MERRA2), and the National Centers for Environmental Prediction-Department of Energy Reanalysis version 2 (NCEP2). The results show that reanalysis temperature (T), specific humidity (Q), geopotential height (GPH), and wind field (U and V components) are consistent with the radiosonde observations in terms of the vertical distribution. In general, ERA5 has the best performance in the CTD during the study period, followed closely by ERA-Interim. However, NCEP2 produces the largest error. The errors of all the reanalysis data show significant diurnal variations, and the diurnal variations differ from each other. Moreover, the results indicate that the reanalysis datasets have largest deviation at 850 hPa (near the ground), which means that the desert region may exist complex interactions between the land surface and the atmosphere. Therefore, more attention should be paid to the description of complex interactions between land and atmosphere over the moving-sand desert region in the numerical models.