Solid biomass combustors are being increasingly deployed for energy supply to help reduce global greenhouse gas emissions, despite the fact that they still suffer from some uncontrolled deflections. Biomass composition variability is a factor that can provoke uncertainty in combustion properties, although it has not been sufficiently studied so far. This paper quantifies the impact of fuel composition variability on thermal and emission aspects of biomass combustion properties in a moving-grate boiler. A one-dimensional transient numerical model of the biomass bed combustion is developed. A set of thermogravimetric analysis experiments on randomly selected biomass pellets are conducted to determine the proximate analysis of the particles. The expected mean value and corresponding standard deviation of fuel particle composition are included in the combustion model in order to analyze the boiler operation data under uncertainty conditions. From this study, it was generally concluded that the fuel combustion properties are profoundly affected by char content, more than moisture and volatile matters. High char content creates less uncertainty, whereas the main source of uncertainty arises from moisture variability. Heat generation from the boiler varies up to 6.7 and 10.7% for wood pellets and bamboo chips, respectively. The flame-temperature fluctuation resulting from composition variability is negligible for both fuel cases. The mean ignition rate for wood pellets can deviate up to 9%, and it increases to 11% for bamboo. Eventually, NOx emission from biomass combustion is vigorously influenced by volatile content variability. This study intuitively concludes that using processed biochar in the combustion system not only improves heat production but also limits uncertainty to a great extent.