A grand canonical Monte Carlo-simulation (GCMC) study is presented focussing on the adsorption of CO2/CH4 mixtures in different nanopore models, including pristine mesoporous carbons, carbon foams, carbon nanotubes (CNTs), and nanopore models modified with hydrophilic carboxylic groups. We also report and discuss the selectivity of the different adsorbent surfaces under a wide range of temperature and pressure. Our results show that foam structures have the highest adsorption capacity of all the pristine structures studied because of its special architecture. The selectivity markedly enhanced after modification, especially at low pressures, and modified CNTs are found to have the highest selectivity among all the models tested. The effect of temperature and pressure is evaluated and the change in the selectivity trends of modified nanopore models are in contrast to that of the pristine models. The results suggest that the separation performance in carbon nanopores is greatly affected by the nature of the architecture and the heterogeneity of the materials. These findings could be beneficial in conventional pressure swing adsorption processes and the nanoporous structures could be used as parts of mixed polymer membranes. The results of this work present some guidelines for the designing nanoporous structures in order to achieve optimal separation of CO2/CH4 mixtures.