Abstract: In order to solve problem of easily eliminating effective signal component of mine gas monitoring data by wavelet de-noising, a de-noising method of mine gas monitoring data based on Hilbert-Huang transform was proposed. Original gas monitoring data sequence is decomposed into a set number of intrinsic mode function components by using empirical mode decomposition, and marginal spectrum is obtained through Hilbert transform. Correlation between the original sequence and each intrinsic mode function component is analyzed according to amplitude frequency relationship in the marginal spectrum between the original signal and each intrinsic mode function component, so as to determine and eliminate noise signal sequence. The case analysis shows that characteristics of the gas monitoring data is clear in time scale by empirical mode decomposition, which is good for identifing high frequency noise of the signal easily. The high frequency noise of gas monitoring data is eliminated through Hilbert spectrum analysis, and intrinsic characteristic of the original gas monitoring data is retained, which avoids signal distortion while de-noising is achieved, and maintains authenticity of the gas monitoring data.