This reaserch examines different analytical models based on the ideal case membrane structure that can use to evaluate gas penetration into mixed-matrix membranes (MMMs) loaded with non-partial fillers. Many models predicted CO2 permeance over PSF/FCTF-1(MMMs) and were compared to experimental results. The models were compared using standard criteria for validating models, such as the difference in penetrant permeability between the two phases ( and the absolute average relative error percentage. A comparison of those models was carried out based on the widely used model validation criteria, including a convenient measure of penetrant permeability difference between the two phases and absolute average relative error percent. Based on the typical values of morphological characteristics, it was determined that the following models fitted the data in the best order: Lewis‐Nielsen model< Pal model<Higuchi< Bruggeman model< Chiew and Gland < Maxwell model having AARE% values of 6.79, 8.45, 8.53, 10.23, 13.10, and 14.33, respectively. A scanning electron microscopy (SEM) examination of the cross-sectional image confirmed that the fillers were really ellipsoids scattered inside the matrix. The Maxwell-Wagner-Sillar model and the Lewis-Nielsen model were then used to evaluate the prolate effect, and the optimization curves of maximum packing () and shape factor (n) produced the least deviations. The AAR% variation was determined to be in the order of 0.01n 0.3, indicating the significance of the shape factor parameter in determining the accurate CO2 permeance. Key words: polysulfone, Mixed matrix membrane, the permeability of CO2, theoretical models