This paper presents a computational study to evaluate the influence of turbulence and combustion models as well as chemistry schemes on the combustion of a 8-11 kW small lab-scale biomass furnace. The analysis is conducted in the zone above the bed (freeboard) where the volatiles are burned. The turbulence models tested are standard k-epsilon, RNG k-epsilon and Realizable k-epsilon; and the combustion models are SFM (Steady Flamelet Model), UFM (Unsteady Flamelet Model) and EDC (Eddy Dissipation Concept). In addition, several chemical mechanisms with different complexity (reduced and detailed chemical kinetics) are considered. The predictions of the velocity, species, and temperature fields are compared with their counterparts' experimental measurements. The present findings reveal that all tested combustion models (SFM, UFM and EDC) are capable of predicting temperature and major species profiles; whereas only EDC is able to reliably predict slow-chemistry species. (C) 2017 Elsevier Ltd. All rights reserved.