The main obstacle of implementing numerical simulations for the prediction of nucleation and epitaxial growth is the variety of physical processes with a considerable difference in time and spatial scales. During the growth of nanostructures and epitaxy deposition of atoms, surface and bulk diffusion, nucleation of two-dimensional and three-dimensional clusters, transitions from two dimensional to three dimensional growth, stress relaxation occur. Thus, it is challenging to describe all of them in the framework of a single physical model. In the present work a multi-scale simulation of the epitaxial growth of silicon carbide nanostructures on silicon using three numerical methods, namely Molecular Dynamics, kinetic Monte Carlo, and the Rate Equations was implemented. Molecular Dynamics was used for the estimation of kinetic parameters of atoms and stress fields at the surface, which are input parameters for the other simulation methods. Kinetic Monte Carlo simulations allowed investigating basic nucleation processes and the transition from two dimensional nucleation to three dimensional cluster growth as well as the ordering of nanoclusters. Furthermore the influence of impurities on the nucleation of nanoscale SiC was studied. The energy barriers values obtained in Molecular Dynamics and the physical model used in the rate equation simulations was validated by Kinetic Monte Carlo. The Rate Equation simulation allowed studying the growth process at larger time scales taking into account the surface stress fields. As a result, a full time scale description spanning over a large substrate area of the morphological and structural surface evolution during SiC formation on Si was developed.