As the latest generation of geostationary meteorological satellite in our country, a great development has been made for Fengyun-4A(FY-4A). Compared with the previous generation(Fengyun-2),FY-4A has higher observation accuracy and shorter scanning time. Taking full advantage of AGRI data, the level of weather and meteorological disasters forecasting in countries along the " The Belt and Road Initiatives" will be effectively improved. The interface for the FY-4A AGRI (Advanced Geosynchronous Radiation Imager) data assimilation is complemented in WRFDA v3.9.1 (Weather Research and Forecasting model’s Data Assimilation v3.9.1) before investigating the bias’s characteristics of FY-4A AGRI data based on RTTOV v11.3 and GFS analysis. Bias-correction experiments of FY-4A AGRI data in infrared channel 8-14 are further conducted. The results show that: 1) Channel 8-10 and 14 have warm biases. There are cold biases in channel 11-13. Because the biases and standard deviation of water vapor channels 9 and 10 are small. The characteristics of biases show obvious differences between land and ocean in channels 11-14. The land’s biases are more complex than ocean’s. For these channels, observations on land can be eliminated in quality control. 2) The slope of linear regression equation between bias and satellite zenith angle is less than 0.035. There is no obvious dependence of biases on the satellite zenith angle.3) The bias in channels 8 and 11-14 show more obvious dependence on scene temperature than those in channels 9-10. 4) The variational bias correction experiment during 1800 UTC 13-15 May 2018 shows that the systematic bias has been corrected effectively.