Aiming at the problem of noise reduction of electrostatic signals in the aero-engine gas-path under strong background noise， a noise reduction method based on intrinsic modal function （IMF） adaptive filtering combined with wavelet thresholding is proposed. Firstly， the original electrostatic signal is decomposed by the complementary ensemble empirical mode decomposition （CEEMD） method to obtain several smooth IMFs. Secondly， the optimal reconstruction adaptive low-pass filtering algorithm is constructed to filter the signal-dominated IMFs. Thirdly， the noise-dominated IMF components are noise-reduced and reconstructed with the signal-dominated IMFs by the wavelet thresholding method， then the noise-reduced electrostatic signal is obtained. The simulated and measured signals are used to verify the proposed method and compare it with other noise reduction methods， and the results show that the method is effective in noise reduction of engine gas-path electrostatic signals and is superior in extracting weak fault signals.