Abstract: At present, seismic waves have a wide range of applications in geological structure detection and structural damage detection. There are various kinds of noise interference during the seismic wave acquisition process, which affects the data processing of the signal. In this paper, the optimal noise reduction model is established by EMD and EEMD algorithms. The noise reduction effect is analyzed and compared by low-pass filtering and band-pass filtering algorithms. The smooth models of the signal are also compared by taking different values for the smoothness coefficient α. In more detail, firstly, the basic principles of EMD and noise reduction smoothing model are introduced. The Gaussian white noise interference is then added via simulated Rayleigh wave; after that, the data is processed by the optimal noise reduction smooth model. Finally, the feasibility of the model is verified through the actual seismic wave. The results show that the EEMD model is effective in the practical application of seismic wave noise reduction.