The frequency of an internal synthetic reference signal that is inconsistent with that of the noise signal is a serious problem in narrowband active noise control (NANC) system. This is called frequency mismatch (FM), which severely deteriorates the performance of the system. To alleviate this problem, a multichannel feedback NANC system with a coarse frequency estimator and spectrum shifter is proposed in this paper. Rough estimations of multiple tones are provided by a coarse frequency estimator based on discrete Fourier transform (DFT) coefficients and fed into the spectrum shifter. Splitting the noise signal into several individual orthogonal reference signals by the spectrum shifter rather than internal generation can avoid FM problem. Because estimated frequency errors can be tolerated in the spectrum shifter but cannot be compensated by signal generators. Compared with existing methodologies, the proposed method relaxes the demand on the precision of the frequency estimator, and reduces the undesired delay in the frequency estimation by using smaller fast Fourier transform (FFT) blocks. Moreover, a computational complexity analysis of the proposed method indicates that the proposed method is advantageous in terms of computational complexity. Extensive simulations and real-time experiments were conducted to verify the superior noise reduction performance and faster convergence speed of the proposed method under different FM situations.