The Atmospheric Infrared Sounder (AIRS) retrieval data include temperature and humidity profiles. The local ensemble transform Kalman filter (LETKF) was developed with the Weather Research and Forecasting (WRF) model, and the impact of assimilating AIRS retrievals is investigated with the WRF-LETKF system. In order to assimilate AIRS data effectively, the adaptive inflation method was improved. In the 6-hourly data assimilation cycle, the four inflation fields at 00, 06, 12, 18 UTC are defined separately. Assimilating AIRS retrievals generally improved longer-lead forecasts. Moreover, quite significant improvement was obtained in the forecasts of Typhoon Sinlaku (2008) in terms of both track and intensity.