Porosity is one of the most important characteristics of fabrics that dictate the permeability and retention properties of fabrics. Several technical uses require textiles with a combination of definite permeability and retention properties. Besides filtration, surgical textiles require these contrary properties to offer an effective barrier against particle laden fluids, such as bacteria and viruses, together with added wearer comfort. Pore size and pore size distribution are important characteristics to determine the permeability and retention behaviour of multifilament barrier textiles by influencing the effective porosity, which can be tailored according to end use requirements by material, weave construction and processing factors. The present research was aimed at developing the relationship that material, construction and loom parameters have with porosity in terms of the mean pore size and mean flow pore size of the fabric, and thereby with air permeability. To map such nonlinear complex relations, an artificial neural network (ANN) was employed. From the findings, it was observed that the porosity of barrier fabrics can be predicted with excellent accuracy using an ANN.