Abstract: Taking falling coal gas emission and gas emission in goaf in the advancing process of working face as the research objects, depending on the characteristics of the two gas emission characteristics, a time series decomposition method for gas emission is proposed. The characteristics of gas emission in goaf, the characteristics of falling coal gas emission and the trend of average gas emission in working face are analyzed. It is considered that the three gas emission characteristics are all relevant and independent. Based on the method of time series decomposition, the characteristics of falling coal gas emission and gas emission from goaf were extracted and analyzed. Taking the data of gas emission of 15 days in Shanxi Province as an example, this paper validates the time series decomposition, feature extraction and analysis method of gas emission. It is considered that the method is effective and feasible.