The precise knowledge of the experimental spectroscopic line parameters of methane is widely used in atmospheric science and astronomical exploration, especially its collision broadening and temperature-dependent coefficients, which are important to study the methane molecular concentration profile. Precise experimental measurement is the significant method to obtain accurate spectral line parameters. In order to derive unknown parameters through a non-linear least-squares method, it is necessary to the simultaneous treatment of several experimental spectra in the same region were recorded with known experimental conditions (concentration, temperature, total pressure, absorption optical path and mixture ratio of gas molecule species, etc.). Then these spectra were usually fitted separately using spectrum-by-spectrum fitting procedure, but this method was often time-consuming and errors were easily produced. In order to solve this problem, a multispectrum fitting procedure based upon the non-linear least-squares and Levenberg-Marquardt method have been developed, able to treat simultaneously several laboratory spectra obtained by tunable diode laser absorption spectroscopy. This procedure can obtain a set of spectral parameters based on the global fitting method. The theory, data processing and application of the procedure are presented in detail. The laboratory spectra of 6 transition lines of ¹²CH₄ molecules in the 2 958~2 959 cm^-1 region are processed by using the Voigt profile of the multispectrum fitting program, and these lines air-broadening coefficients were obtained at five temperatures (296.0, 251.0, 223.0, 198.0, 173.0 K). The air-broadening coefficients of the present work are compared with the available values reported in the literature, which were obtained by spectrum-by-spectrum fitting procedure, and the difference between them lies within -4.97%~1.58%. The results show that the broadening coefficients of these transitions in this work agree well with those available values. Among the 30 sets of correlation data, there are 4 sets whose fitting errors obtained by the spectrum-by-spectrum fitting program are smaller than those by the multispectrum fitting program, 2 sets with equal error values, and the remaining 24 sets whose fitting errors obtained by the multispectrum fitting program are smaller. Therefore, the multispectrum fitting program has good reliability and is suitable for processing gas molecular absorption spectrum.