A current challenge of physical, chemical and engineering sciences is to develop theoretical tools for predicting structure and physical properties of hybrid organic inorganic nanocomposite from the knowledge of a few input parameters. However, despite all efforts, progress in the prediction of macroscopic physical properties from structure has been slow. Major difficulties relate to the fact that (a) the microstructural elements in multiphase material are not shaped or oriented as in the idealizations of computer simulations, and more than one type can coexist; (b) multiple length and time scales are generally involved and must be taken into account, when overall thermodynamic and mechanical properties wish to be determined, and finally (c) the effect of the interphases/interfaces on the physical properties is often not well understood and characterized. As a consequence, their role is often neglected in the development of new theoretical tools or they are treated in a very empirical way. In this work, we focused on issues (b) and (c) in a multiscale molecular simulation framework, with the ultimate goal of developing a computationally-based nanocomposite designing tool. In particular, we developed a hierarchical procedure in which lower scale (i.e., QM, MD and /or MC) simulations are performed to obtain parameters for higher scale (i.e., mesoscopic and/or finite element) calculations, from which the bulk properties of the hybrid nanocomposite material can be ultimately estimated.