Abstract: As there are interference sources in the surroundings and instruments for the collections of the seismic wave signals, the seismic waves collected may always be interrupted by many noises, which affacts the further applications of the seismic waves. The mode cell filter is a novel adaptive denoising method with no need for pre-defined basic functions. In this paper the MCF is introduced to the seismic wave denoising, and a MCF-based automobile noise eliminating method is proposed. Experiments on artificial and natural signals are designed for denoising. The results show that, the performances of MCF-based denoising method are better than the selected optimal wavelet threshold method in the artificial case, while in the natural case, the MCF-based method is better than IIR digital filter and close to the improved optimal wavelet threshold method.