Efficient scale-up processes of gas separation membranes require a thorough assessment of performance across various parameters. Semi-empirical modeling based on theory and empirical data is an attractive approach to predict the required performance. Despite numerous studies on the performance of gas separation membranes depending on internal parameters such as morphology and geometry, relatively few studies on performance dependent on external parameters have been reported. In this study, we propose semi-empirical modeling for the performance of membranes as a function of external parameters using a bench-scale pore-filled composite membranes for water vapor separation. As external parameters, we selected temperature, pressure, relative humidity (RH), and feed flow rate. These parameters influence water vapor permeance, water flux, and RH reduction, which are theoretical indicators of the performance of water vapor separation membranes. A power regression analysis was used to find the power of individual parameters (Y = Coefficient (k) ∙ Xpower), and a multiple linear regression analysis (finite power law model, Y = k ∙ X1α ∙ X2β…) was used to develop a semi-empirical model with the correlations of these parameters. The modeling values closely coincided with measured values, suggesting the effectiveness of the proposed modeling. This method is expected to simplify the prediction of membrane performance, offering valuable insights for potential scale-up processes.