In active noise control systems, the Filtered-x Recursive Least Squares (FxRLS) algorithms has a fast convergence speed, but large computation cost. In this paper a combination method using the lattice structure of joint-process estimation and QR-Decomposition-based Least Squares Lattice (QRD-LSL) active filter algorithms is proposed to control unwanted noise. The proposed method develops the joint-process estimation and obtains the new weighted coefficients update relationship based on the individual error estimation. The Lattice structure of joint-process estimation is simple. The QRD-LSL active filter algorithm has good numerical properties. Simulation results show that the proposed method has good noise control performance, such as fast convergence speed, short computation time, small steady-state error and good tracking capability.