The fast development of digitalisation and computational sciences is opening new possibilities for the rapid design of new materials. Computational tools coupled with focused experiments can be successfully used for the design of new nanostructured materials in different sectors, including membrane engineering. Accordingly, in this chapter, we present the application of a multiscale molecular simulation protocol for predicting gas transport properties in polymeric nanocomposite membranes constituted by titania (TiO2) nanoparticles dispersed in thermoplastic polyurethanes (TPUs). The chapter starts with a general introduction on the future of computational tools for the design of new materials and introduces the underlying paradigm of multiscale molecular modelling. It then continues with the description of a multiscale (i.e., atomistic, mesoscale, and finite element calculations) computational recipe developed ad hoc for the prediction of gas permeation and diffusion in TPU/TiO2 nanocomposite membranes. Finally, a comparison of the in silico and experimental results on these systems is reported and discussed. The quality of the agreement obtained between virtual and real data for such complex systems indeed confirms the validity of such computational tools for the design and transport property prediction of nanocomposite membranes for gas treatment.