In view of the problem that high frequency noise usually mixes in the process of microseismic signal acquisition in mines.The noise reduction method based on the complementary ensemble empirical modal decomposition (CEEMD) and permutation entropy is proposed for noise reduction of microseismic signals.The signal is decomposed into several eigenmode function from high to low frequencies using CEEMD;the ranking entropy value of each mode function component is calculated to quantitatively characterize its degree of randomness;wavelet threshold noise reduction of the IMF component containing more noise information is conducted and the noise-reduced microseismic signal is obtained by signal reconstruction.The noise reduction effect of simulated and real microseismic signals shows that the noise reduction effect based on CEEMD and permutation entropy is better than that of CEEMD and EMD noise reduction method,which can effectively remove the high frequency noise in microseismic signals and provide technical support for further analysis of microseismic signals.