Using the L-band sounding data from Shanghai Baoshan Station, the inversion of temperature data of GIIRS/FY-4A Geostationary Interferometric Infrared Sounder (GIIRS) with different quality control data was analyzed under clear sky, cloudy sky and all sky in typhoon season within 24 hours before and after typhoon landing. The results show that: (1) The inversion accuracy of GIIRS temperature is best under clear sky, and the data root mean squared error (RMSE) with quality control of 0 is 1.74 K, which indicates that the inversion product in the upper troposphere has a certain degree of credibility. (2) The cloud layer reduces the accuracy of GIIRS temperature inversion. The data RMSE with quality control of 0 is 3.57 K under cloudy sky, which exceeds the standard error range given by the US Environmental Monitoring System. (3) When the temperature is greater than 230 K under cloudy sky, the inversion temperature of GIIRS is lower than the sounding data detection temperature. (4) Within 24 hours before the landing of typhoon “Ampil” and “Jongdari”, at the tropopause layer and from the height of 500 hPa to the planetary boundary layer, GIIRS inversion temperature is higher; within 24 hours after the landing of typhoon “Jongdari”, from the height of 800 hPa to the phanetary boundary layer,the inversion temperature of GIIRS is lower. The inversion will generate a large number of invalid values.