Hybridization of metal organic frameworks (MOFs) with graphene oxide (GO) is used to improve the CO2 adsorption performance of MOFs, but the underlying mechanism of this process is still unclear. This study provides a general framework to understanding the mechanism of CO2 adsorption and separation on GO/CuBTC and GO/UTSA-16 in order to optimize the synthesis of the desired material. For this purpose, molecular models mimicking the experimentally available hybrid materials were developed and studied by molecular simulations. Once the models were validated with available experimental data, a systematic study on the effect of different structure variables was performed, searching for the best hybridization procedure for this application, in a predictive manner. It has been confirmed that the interface between GO and MOFs produces strong interactions with CO2, which, together with the smaller pore sizes, significantly enhances the adsorption performance at low pressures. Moreover, the performance of the most promising hybrid GO/MOFs structures from pure CO2 adsorption isotherms for separating CO2 from nitrogen were predicted by GCMC based on binary mixtures (15CO2:85 N2) and a temperature swing adsorption (TSA) process. Among the different materials/compositions explored, GO/CuBTC with the highest GO content (i.e., 65% wt.) and under the premise of no stacking of GO, shows the best results in terms of key performance indicators: CO2/N2 adsorption selectivity (120 at 313 K), working capacity (1.794 mmol/g at a desorption temperature of 443 K), and a specific energy consumption (0.534 GJ/tonne-CO2) comparable to amine scrubbing.