In order to contribute to the agricultural residue recovery in a sustainable way in Colombia, an atmospheric bubbling fluidized bed reactor fed with rice husk (with nominal capacity of 150kWth) was evaluated under gasification conditions. This system was assessed using a statistical design of experiments based on the response surface methodology-RSM (central composite design-CCD) in order to obtain empirical correlations for describing the performance behavior of the process in terms of the equivalence ratio (ER) and the normalized fluidization velocity (Unc). The lower heating value, the volumetric gas yield, the gas power and the cold gasification efficiency were the output variables. Optimum conditions with respect to both the lower heating value (3.78MJ/Nm3) and the gas power (73.82kW) were obtained at ER of 0.24 and Unc of 0.19m/s. However, the height of the initial fixed bed of inerts requires to be higher in order to increase the mass and energy transfer conditions and hence the carbon conversion. The statistical correlations obtained are considered acceptable taking into account the complexity of the gasification phenomena in fluidized bed. The results are considered as a base for designing small scale decentralized power generation units using rice husk as feedstock.