The molecular simulations (Grand canonical Monte Carlo (GCMC) and molecular dynamics (MD) simulations) combined with ideal adsorbed solution theory (IAST) are adopted to investigate the adsorption of CO2, CH4 and their mixture in multilayer graphene nanostructure. The effects of pressure, temperature and pre-adsorbed water on the separation behaviors are examined. The IAST accurately predict the loading of two species, but it has a slight deviation for the selectivity predictions. It is beneficial to the CO2/CH4 mixture separation by reducing temperature or pre-adsorbing some water. Due to additional adsorbate-H2O interactions, the diffusivities of two species drop down as the pre-adsorbed water content increases. (C) 2016 Elsevier B.V. All rights reserved.