Biomass burning is an essential contributor to smoke concentrations. It is challenging to find the optimal fire emission inventory to estimate fire emissions, that are crucial for predicting air quality. This study aims to investigate the effect of two global biomass burning emission inventories, the Global Fire Emissions Database 4 (GFED4) and the Quick-Fire Emissions Database 2.5 (QFED2.5), on simulating PM2.5 concentrations in Canada. The two inventories were assessed using GEOS-Chem modelling system for the 2020 fire season in Canada (May-August). The simulated PM2.5 concentrations illustrated slight differences at the six air quality monitoring stations and better model performance with QFED2.5. Additionally, the results of sensitivity analysis showed that the Canadian biomass burning led to an increase in the concentrations of PM2.5, more than 42% at Fort Smith and Yellowknife stations. In conclusion, this study suggests that QFED2.5 is more suitable for simulating PM2.5 in the assessed regions.