In this dissertation, we present a multi-scale simulation method for the efficient calculation of semiconductor nanostructures that are too large to handle fully first-principles calculations. Firstly, recently an effective mass approximation method based on atomistic density functional theory calculations results such as dielectric constants, effective mass, Kohn-Sham potential was published. Here, we report a grid-based multi-scale simulation method applied to the prediction of optical gaps in quantum dots, nanorods, and nanoplatelets by further expanding the first-principles-derived effective mass approximation. Furthermore, we developed a grid-based adaptive mesh refinement for multi-scale simulation We anticipate that this study will become a cornerstone for multi-scale simulations that are difficult to apply fully first-principle calculations.