The microseismic signals collected in the process of mine microseismic monitoring are mixed with interference signals. It will seriously affect microseismic signal recognition, first arrival picking and source location. To accurately obtain effective microseismic signals from the original collected signals, a denoising method based on novel adaptive ensemble empirical mode decomposition (NAEEMD) and frequency constrained singular value decomposition (SVD) is proposed. Firstly, the original mixed signal is decomposed into intrinsic modal components (IMF) with the order from high to low and residual components by NAEEMD. Then, the transition component is determined according to the correlation coefficient, and the IMF components are adaptively divided into the signal-dominated IMF components and the noise-dominated IMF components. Taking the sum of the signal-dominated IMF components as the frequency constraint condition, the noise-dominated IMF component and transition component are denoised by SVD. Finally, the denoised components and residual components are reconstructed to obtain the denoised microseismic signal. The simulation analysis is carried out with the simulated signal. The results indicate that the proposed method is easier to obtain useful signals and has good frequency convergence and signal-to-noise ratio (SNR). In addition, the denoising experiment is carried out with the measured signal. Compared with empirical mode decomposition (EMD), ensemble empirical mode decomposition (EEMD), SVD, and NAEEMD, the three evaluation indexes of SNR, energy percentage (E), and standard deviation (RMSE) are calculated quantitatively. The results show that the SNR of the proposed method is 15.6258 dB higher than that of other methods, the E is as high as 94.8625%, and the RMSE is 0.0216. The proposed method is effective in denoising mine microseismic signals and has better denoising effect than EMD, EEMD, SVD, and NAEEMD. [ABSTRACT FROM AUTHOR]