The NOAA National Centers for Environmental Prediction (NCEP) Global Forecast System/Global Data Analysis System (GFS/GDAS) is the decision support system used by NOAA for medium-range numerical weather predictions. The forecast model, GFS, is a global spectral model with meteorology data assimilation using the Gridpoint Statistical Interpolation (GSI) scheme. Climatology-derived aerosol distributions are currently used in the GFS radiation module and background aerosol conditions are assumed in the GDAS radiative transfer scheme. Consequently, the effects of aerosols on radiation, clouds, and convection are poorly constrained in the GFS and the effects of aerosol attenuation on radiance are yet to be quantified in the GDAS. The NCEP Environmental Modeling Center (EMC) is developing an aerosol forecasting and assimilation capability in GFS/GSI to improve the treatment of radiation feedback in the GFS forecasts and to provide aerosols boundary conditions to the CMAQ-WRF regional air quality. We will overview the progress of the NCEP global aerosol modeling system and the approach on the chemical data assimilation. Results of GFS/GSI experiments for 2006 summer period will be presented and the impact of aerosols on medium range forecasts will be discussed.