Abstract: In order to improve the performance of active noise control effectively, this paper presents a model reference hybrid direct adaptive control(MRHDAC) strategy. Based on the model reference direct adaptive control(MRDAC), this strategy makes use of advantages of FLMS algorithm and nonlinear BP networks to restructure a control system. Weights of FLMS and BP algorithm were modified according to the system error. When it begins to control nonlinear system, the system applies mainly FLMS algorithm to improving the convergent rate, then, BP to improve the control accuracy. The hybrid system keeps the advantages of the BP networks and at the same time improves convergence rate. The results of experiments have proved that the new strategy is more effective than MRDAC.