The use of waste biomass, e. g., sawdust, is an alternative for heating and industry in countries with large areas of radiata pine and eucalyptus forests. Indeed, the Chilean demand for domestic, industrial, and district heating has been increasing in recent years, standing at 75 kt/y and increasing by 400 % until 2030. However, burning hydrophilic sawdust pellets and wooden logs generates particle matter, NOx, and polycyclic aromatic hydrocarbons (PAHs) emissions that have placed this fuel as one of the most contributing pollutants to the atmosphere in southern Chile. According to current emission inventories, residential biomass combustion contributed more than 95 % of total particulate matter emissions. It has become a limitation for biomass as a source of energy and a challenge for applying Waste-to-Energy technologies to obtain biochars with greater calorific value and less ash. On the other hand, PAHs are the most problematic pollutants group to reduce in the solid and gas phase due to their high molecular weight. It is essential to study a system that diminishes those pollutants by improving the solid fuel and the pellet stove and developing an efficient pollution control system, like a PAHs biofilter. This research studied the emission of benzo(a)pyrene, a model compound of PAHs, during the combustion of commercial pellets in a stove. The PAHs emissions, determined by gas chromatography, were between 17.8 and 22.4 µg/kg of pellets with an operating temperature range of 350 to 450 °C. A biofiltration system treated the exit combustion gas for pollutant abatement. Vermiculite was used as a support medium coated with a microbial consortium composed of Fusarium solanis and Rhodococcus erythropolis. The system showed an efficiency of nearly 70 % and an elimination capacity of 250 µgm-3h-1. On the other hand, for the CFD modeling, the geometry was obtained by a mathematical algorithm of random generation for packed beds in bioreactors. The model meshing and numerical solution were carried out in OpenFOAM, with the solvers SimpleFoam and scalarTransportFoam. The resolution of the velocity-pressure field using SimpleFOAM showed a maximum error of 6.1 %, and for the concentration of the pollutant, the error was 6.2 %.