This study aims to propose the importance of employing statistical analysis and modeling for investigation on the design and optimization of CMS membranes for gas separation. The factorial methodology is used to optimize permeability and selectivity through implementing the general factorial design considering the three main parameters such as choice of precursor materials, blend composition and the final pyrolysis temperature. Findings by statistical analysis showed that the linear and quadric terms of these three variables had significant effects. The optimal conditions are the Matrimid, blend composition 70%, and pyrolysis temperature at 800 degrees C. Under these conditions, the model estimated a CO2/CH4, H-2/CO2, and O-2/N-2 selectivity of 120.2, 8.39, and 8.06, respectively. Employed statistical technique and developed models can be used as a useful tool for design and optimization of appropriate gas separation membranes with effective performance for various industrial applications. POLYM. ENG. SCI., 54:147-157, 2014. (c) 2013 Society of Plastics Engineers