The paper presents the results of the development of a regression model for reconstruction of the total methane content [CH4] in the atmosphere according to the data from AIRS hyperspectrometer of the Aqua satellite. The information basis of the study is the results of the chemical transport model MOZART-4 modified by the authors for the period 2001-2020. It is shown that the proposed model, which accepts only AIRS L2 products as input, provides reconstruction with a correlation coefficient R2 = 0.94 and a standard deviation σ = 6.2 ppb. A comparison of the reconstruction results with the observations of the TCCON network of ground-based stations in Europe showed a satisfactory agreement between the results of the developed model and measurements. Using the proposed method, the rate of change of the total methane content in the atmosphere of Western Siberia in 2003-2022 was estimated, which is equal to 5.5 ± 0.2 ppb/year.