The uncertainty in biomass burning emissions are large in many regions due to high variation of fire characteristics, limitation of fire data and uncertainty in factors calculation. The simulations in Mainland Southeast Asia (MSEA), using Fire Inventory from NCAR (FINN) as the biomass emissions that is estimated by using emission factors, fuel consumption and burned area based on land use/land cover (LULC) type, found great overestimation of PM in dense biomass burning area. This study developed the method to solve the uncertainty of biomass burning emission by air quality modeling system using Fire Radiative Power (FRP) to modify FINN. The modeling system WRF-CMAQ is applied to simulated PM in MSEA during smog episode in March 2012. Results from simulation were compared to both satellite and ground-based observations. The comparison of simulated PM with modified FINN by FRP (PMFINN-FRP) showed generally good agreement and the modeling system captured most of the important observed features. The comparison of PM in source region found greatly improvement of simulated PM. Simulated PMFINN-FRP are in factor of two of the observations more than 70% and spatial correlation with the observation are greater than 0.8.