For most polymeric membranes, the gas permeability coefficient (P) is often interpreted as the product of diffusivity (D) and solubility (S) of a penetrant gas in the polymer (P=D S). The basic assumption is that molecular diffusion is primarily responsible for mass transport in the membrane permeation process. However, for some open structure membranes, such as poly(1-trimethylsilyl-1-propyne) [PTMSP] or poly(dimethylsiloxane) [PDMS], the high permeabilities of some gases yield much higher diffusivities when calculated from the above relationship (P=D S) than when calculated by using the direct kinetic measurement of diffusivity. It is hypothesized that this discrepancy is due to the convective transport of gas molecules through such open structured polymers. In most cases, the convective contribution to mass transport through membranes is negligible. However, for polymer membranes with high free volume, such as PTMSP, whose free volume fraction is 20 to 25%, the convective term may dominate the permeation flux. In this study, a non-equilibrium thermodynamic formalism is employed to properly treat the diffusion term and convective term that constitute the Nernst-Planck equation. The current analysis indicates that the total permeation flux, which consists of a diffusion term and a convective term, agrees well with the experimental data for several permeation systems: pure components propane and n-butane/PTMSP, pure gas hydrogen/PTMSP, and mixed gas hydrogen/PTMSP. Also, the permeation systems of a nonporous rubbery membrane, PDMS, and eight organophosphorus compounds were included in the study. It is recommended that the proposed model be validated by using other polymers with high free volumes and high permeabilities of gases and vapors, such as poly(1-trimethylgermyl-1-propyne) [PTMGeP] and poly(4-methyl-2-pentyne) [PMP].