Abstract: There exists noise signal in the kinetic experiment of desulfurization of flue gas with rare earth oxides by means of isotherm thermogravimetry because of experimental conditions. The contributing factors of the denoising of thermogravimetric signal were discussed by the Daubechies (N=3) function of wavelet transform and Matlab software. The results showed that different rules of threshold and methods of threshold have little effect on denoising, but the number of decomposition layers play an important role in denoising. The optimal number of decomposition layers could be obtained when the rate of signal to noise changed slow. The signal represented the kinetics process of the isotherm thermogravimetric experiment truly after denising.