Pore network modeling of porous media has this advantage that can consider the pore structure incorporating any desired details, but it has not been studied sufficiently. In addition, most studies are limited to mathematical modeling only which need validation. In the present study, this approach was applied to hydrogen separation from syngas by nanoporous ceramic membrane to predict the membrane permeance theoretically based on its pore structure. Gas transport through nanoporous membrane was modeled with the aim of a 2D network model. A dusty gas model was used for gas transport in the individual pores. Model validation showed that the model predictions are in good agreement with the experimental data with the coordination number of 2.5 and the pore length of about 20nm. A parametric study indicated that hydrogen permeance through the membrane increases with the average and minimum pore size and decreases with temperature and pressure. Also, hydrogen selectivity increases slightly with temperature and decreases with pressure and average pore size.