Abstract: Combined with porous media model and RNG k-epsilon turbulent flow model，numerical model for air flow in hot air furnace of clean coal reverse combustion for dense flue-cured tobacco had been established.Variation of air flow distribution and air utilization rate in furnace with height of coal bed，height of clean ash on side wall of static pressure area and degree of alignment of honeycomb holes of upper and lower briquettes were studied by using single factor simulation optimization method.Results showed that air utilization ratio was cubic function of air velocity，cubic function of coal bed height，quadratic function of clean ash mouth height and quadratic function of resistance coefficient in porous media area.Precision of controlling heat supply of reverse combustion furnace was independent of viscous resistance coefficient of porous media.Preloading proper amount of briquetting coal，pile-to-hole stacking briquettes and closed ash cleaning nozzles were some preconditions for accurate heating control of reverse combustion furnace.This study provided a theoretical basis for design optimization of reverse combustion furnace.