The disturbance picked up by error microphone can significantly degrade the steady-state performance of active noise control (ANC) systems. Previously a cascading algorithm has been proposed, where a supporting adaptive filter is used for removing the uncorrelated disturbance from the error signal of filtered-x least mean square (FxLMS) algorithm. We modify this algorithm to improve the performance in terms of fast convergence speed and reduced steady-state fluctuations. We modify the desired response of the supporting filter which gives us better control to tune its parameters. This modification allows to use modified FxLMS (MFxLMS) algorithm to adapt ANC filter, which results in improved convergence speed. Computer simulations are carried out which demonstrate the effectiveness of the proposed method. This improved performance is achieved at an increased computational complexity.