The FxLMS is widely used technique in active noise control. In conventional FxLMS, value of convergence coefficient is kept constant which may not be optimum at all frequencies. At some frequencies, convergence speed of FxLMS is slow and at some, it is relatively fast. This leads to significant degradation in performance of FxLMS if characteristics of noise are continuously varying. Normalized FxLMS algorithm is not very effective for continually varying noise. Eigenvalue equalization method may become unstable at some frequencies. Other methods like FxGAL are computationally expensive. The method proposed in this paper attempts to optimally adapt the convergence coefficient of FxLMS algorithm for continually varying noise. It is based on estimating how frequency of noise is varying using fast Fourier transform of reference signal and then, using this information to optimally adapt the convergence coefficient. The optimum value of convergence coefficient depends upon power and phase delay of filtered reference signal. A numerical study of a 3D acoustic cavity shows that the proposed method leads to faster convergence giving higher noise reduction. It also shows that higher the rate of variation of frequency of primary noise, higher is the advantage of the proposed method over the conventional FxLMS algorithm in terms of noise reduction. © 2015 by ASME.