It is difficult to eliminate the effect of interfering signals on automobile wheel speed signal by the traditional processing method. According to signals and noises presenting different characteristics at every scale space after wavelet transform, this paper applied wavelet theory to decompose the noise signal for several levels by selecting Daubechies(db4) wavelet. Aiming at every level detail signal after wavelet decomposing, the soft-threshold denoising method was employed to eliminate noise, then the inverse wavelet transform could be used to reconstruct signals for signal denoising and rebuilding. The researching results show that this method can effectively eliminate the noises of wheel signals.