An experimental and statistical study was performed for the purification of hydrogen in an H2/CO2 mixture using the PSA process with activated carbon. This process has been extensively used for the high purification of gas mixtures. Central composite design (CCD) was used for the modeling and optimization of the process, regardless of complex and time-consuming equations generally employed in the literature. The statistical analysis of models, as obtained from the CCD method, revealed that distinct second-order polynomial equations, with F-value more than 200, p-value less than 0.0001, and R-squared more than 0.9900, could predict the experimental data for purity, recovery and productivity of hydrogen. Based on the established models, as the purge-to-feed (P/F) ratio increased, the purity increased, while the recovery and productivity decreased. The recovery and productivity first increased and then decreased as adsorption time increased. Thus, the optimal condition of the PSA process was obtained using the CCD models.