Abstract: In order to reduce the impact of trend and noise components on the accuracy of tunnel blasting vibration signal processing, a preprocessing method for blasting vibration signals based on Fourier Decomposition Method (FDM) is proposed. Firstly, the measured blasting vibration signal is decomposed into a series of components through FDM, and the signal dominant component is obtained based on the principle of threshold selection. Then, the dominant components are reconstructed to obtain a pure signal after removing trend terms and noise, and the effectiveness of the method is verified using spectral analysis. The research results show that FDM can effectively separate low-frequency trend terms and high-frequency noise in blasting vibration signals. Compared with existing traditional methods, the pure signal time history curve obtained after FDM processing is the smoothest, with the highest signal-to-noise ratio (23.240 6) and the smallest root mean square difference (0.024 6), demonstrating good preprocessing performance.