Using the observations made by the hyperspectral infrared atmospheric sounder (HIRAS) carried on the FY-3E polar orbiting meteorological satellite, the back propagation neural network (BPNN) method is employed to conduct research on the inversion of atmospheric temperature and relative humidity vertical profiles in the East China region. An atmospheric temperature and humidity inversion model is constructed, and the parameters are optimized to obtain a network model configuration with high inversion accuracy, resulting in all-weather, high-precision atmospheric temperature and relative humidity profile. According to the results, the following conclusions can be drawn. 1) Temperature inversion results of the model have a mean error (ME) between -1.00 K and 1.00 K at each pressure layer, except that the absolute value of ME in the lower layer of the cloudy sky is greater than 1.00 K. The validation experiment conducts on ERA5 data has root mean square error (RMSE) between 0 K and 2.00 K for most pressure layers, except for ~2.50 K at the 925?950 hPa layer, the minimum RMSE corresponds to the 200?500 hPa, indicating higher temperature inversion accuracy in the upper atmosphere. 2) For the inversion of atmospheric relative humidity, the ME of the lower and upper atmosphere is larger, whereas that of the middle atmosphere is smaller; RMSE is larger in the middle layers and smaller in the lower and upper layers. 3) Compared to clear sky condition, the accuracy of temperature and relative humidity inversion models under cloudy sky condition is slightly lower. 4) The deviation of the temperature and relative humidity inversion results from the sounding data is slightly greater than the deviation from the ERA5 data although the trend of error with height variation is similar. The HIRAS data generally performs well in inverting the temperature and relative humidity of clear and cloudy skies, with high inversion accuracy. Therefore, this study has important reference value for inversion methods and techniques of atmospheric temperature and relative humidity, providing useful insights for future related research.