The TDLAS system combined with multi-path absorption cell is prone to interference noise in practical application, which has a very adverse impact on its detection performance. In this work, the causes of this interference are analyzed, and a wavelet denoising method is proposed to improve the detection performance of the system. Firstly, the appropriate wavelet function and decomposition layers are selected according to the theoretical research results, and then the simulation signal with superimposed interference is filtered through this wavelet. The results show that this noise reduction technology has good application performance. Finally, the direct absorption spectrum (DAS) and second harmonic signal of each carbon dioxide concentration collected in the experiment are processed by wavelet denoising technology. Compared with the original signal, the signal-to-noise ratio of the denoised signal is increased from 0.4 to 259, and the detection limitof the system reaches 7×10-6, which indicates that the proposed small wave noise reduction method has high application value in the detection of ambient gases.