Abstract: Satellite microwave remote sensing data are one of the most important data required by global/regional assimilation systems. The brightness temperatures of the upper atmosphere (middle and upper stratosphere) from microwave remote sensing are affected by many factors and have certain errors. Evaluation of the accuracy of microwave remote sensing of the upper atmosphere has become an important research topic owing to the lack of high-accuracy sounding data of the upper atmosphere. In this study, the accuracy of the observation data of on-orbit microwave temperature radiometers, such as Advanced Microwave Sounding Unit-A (AMSU-A), Advanced Technology Microwave Sounder (ATMS), and FengYun-3D Microwave Temperature Sounder-2 (FY-3D MWTS-2), was examined and evaluated in 2020. Three evaluation methods (comparison with radio occultation observation, radiation transfer simulation, and cross-calibration) were comprehensively used for the analysis. From different perspectives, the three methods revealed the error characteristics of the data in the upper atmosphere from various instruments and the possible sources of the errors analyzed. The specific deviation values of the brightness temperatures of high-level channels differed from those achieved using the three evaluation methods; however, the basic trend of the variation remained consistent, i.e., high-level channels had greater noise than middle- and low-level channels. In addition to the cross-calibration method, the other two methods demonstrated seasonal variation in the deviation. Overall, AMSU-A outperformed FY-3D MWTS-2 and ATMS. The temporal and spatial variation characteristics of brightness temperature accuracy in the upper atmosphere can provide a reference for microwave data assimilation and climate applications.