The unique physical properties of nanomaterials, attributed to the combined effects of their size, shape, and composition, have sparked significant interest in the field of nanotechnology. Fabrication of nanodevices using nanomaterials as building-blocks are underway to enable novel technological applications. A fundamental understanding on the structure-property relationships and the mechanism of synthesizing nanomaterials with tailored physical properties is critical for a rationale design of functional nanodevices. In this thesis, molecular simulations that employ a detailed atomistic description of the nanoscopic structures were used to understand the structure-transport property relationships in two novel classes of porous nanomaterials, namely, polymer/porous inorganic layered nanocomposite materials and single-walled metal oxide nanotubes, and provide predictions for the design of nanodevices using these nanomaterials. We employed molecular dynamics to study transport of gas molecules (in particular He, H2, N2 and O2) through a polydimethylsiloxane/porous layered silicate (AMH-3) nanocomposite membrane material as a function of its composition. Gas separation performance of the nanocomposite was found to be substantially enhanced for H2/N2 and H2/O2 compared to pure polymeric material due to the molecular sieving effect of AMH-3, suggesting the possibility of developing a new class of superior separation devices. We also developed force field parameters for layered aluminophosphates that are emerging as potential inorganic layers for construction of nanocomposite materials. We presented preliminary work on developing Transition State Approach-Monte Carlo simulation method for calculating gas transport properties of nanocomposite materials. We investigated in detail the diameter control phenomenon in single-walled metal oxide nanotubes using molecular dynamics simulations and demonstrated the existence of a thermodynamic 'handle' for tuning the nanotube diameters and derived a unique correlation between nanotube energy, composition, and diameter to precisely predict nanotube diameters. Finally, using a combination of molecular dynamics, monte carlo and sorption experiments, we investigated adsorption and diffusion properties of water in single-walled aluminosilicate nanotubes. We predicted high water fluxes in these nanotubes, due to short lengths, hydrophilic interior and near-bulk-water diffusivities. Overall, my research represents two examples of the progress in developing a predictive basis for the design and analysis of nanostructures for applications in separations, nanofluidics, and fuel cell technology.