The information in spectroscopic databases is routinely updated, therefore it is necessary to evaluate on a regular basis how efficient the improvements in spectroscopic data are with respect to the atmospheric remote sensing problems. One of such problems is the methane total column retrieval from the ground-based FTIR measurements of the direct solar radiation. The CH4 line parameters contained in various modern spectroscopic databases demonstrate some differences. In the present study, the impact of these differences on the atmospheric methane retrievals is estimated. The ground-based FTIR measurements at the St.Petersburg and Kourovka observational stations (Russian Federation) are taken as the basis for the study. The retrieval algorithms conform to the specifications worked out by the IRWG NDACC and TCCON networks. The databases HITRAN2001, HITRAN2008, HITRAN2012, HITRAN2016, GEISA, ATM12, ATM16, ATM19 and the GOSAT2014 line list are used as the input for the forward modeling and for the retrievals of the CH4 total column. Different criteria for quality assessment of fitting of measured and simulated transmittance spectra are considered. It has been shown that the variability of the methane total column values retrieved using absorption line parameters from different spectroscopic databases reaches 1.7%. The obtained results confirmed that for mid-infrared range the current IRWG NDACC recommendation to use HITRAN2001 is still valid: the HITRAN2001 data provide better CH4 retrievals than the newer spectroscopic databases, with the exception of one database (ATM19) yielding slightly different results but comparable to HITRAN2001. The best retrieval results in the near-infrared region are observed when the ATM, HITRAN2008 and GEISA2011 databases are used. The lowest value of RMS corresponded to the calculations which used the ATM19 line list.