Abstract:ATOVS data make up for the conventional observations in some areas with scarce stations, and can particularly provide the temperature and humidity data under the condition with cloud for precipitation forecast by model. In order to examine the application and capacity of T213 SSI system and evaluate the impact of ATOVS data on numerical forecast of heavy rainfall in China, the T213 SSI system is used to assimilate ATOVS data from NOAA16, 17 satellites. Then a heavy rainfall event during July 7 to 9, 2007 over the Huaihe River Basin is simulated with different assimilation schemes in order to compare and analyze the results of assimilation and simulation. The main conclusions to be drawn are as follows: after AMSU data being added into the T213L31 model, the forecast of precipitation is more exact, especially the intensity of precipitation. By continuous assimilation, temperature fields and humidity fields are improved obviously. After eliminating AMSU A terrestrial channel data, the simulating effect is better than assimilating completely on the location and the intensity of the heavy rainfall centre. Especially, the synoptic situation, the temperature, humidity and wind fields are improved much closer to the real states compared with the assimilating AMSU A terrestrial channel data.