The impact of MODIS retrieval products of the temperature and humidity profiles in the data assimilation and numerical prediction of tropical cyclone (TC) are discussed in this paper. The retrieval products are derived using the MODIS Collection-5 algorithm and data from NASA EOS AM (Terra) and EOS PM (Aqua) satellites collected in real-time. The Collection-5 software generates temperature and relative humidity on 11 pressure levels from 1000 hPa to 100 hPa over clear to partially cloudy regions. Using radiosonde data over southern China, verification of the MODIS profile products were conducted to study the error characteristics. Numerical experiments using the JMA Non-Hydrostatic Model (NHM) and its 3-dimensional data assimilation system (3DVAR), configured at horizontal resolution of 10 km, were carried out to investigate the effects of the retrieved profiles in model analysis and forecast for selected TC cases in 2009 and 2010. It is found that the MODIS retrieval products help to improve the moisture and temperature fields in the initial condition of NHM and enhance the precipitation development associated with the TC circulation. Improvements in the track and intensity forecasts can be seen through the ingestion of MODIS profiles in the 3DVAR system of NHM.