Seismic hazard analysis has progressed over the years due to the advances in the availability of seismicity and ground motion data of damaging earthquakes. In the seismic hazard modelling community, source mechanism-based models are becoming popular to simulate ground motions. This mechanics-based model is useful as it takes into account the intricacies of the rupture process, material property variations, basins, ridges, bathymetry, and topography. The present thesis explores physics-based ground motion simulation in seismic hazard analysis, specifically to study the seismic response of basins. For this, a detailed 3D computational model of India, by incorporating the intricate details is developed in spectral element framework. The applications of the developed computational model are demonstrated for Indo Gangetic (IG) basin, Kutch basin, and Kathmandu basin by simulating ground motion for past damaging earthquakes. The developed computational model is validated qualitatively and quantitatively using the records of the 2001 Mw 7.6 Bhuj event (for Kutch basin) and 2015 Mw 7.9 Nepal event and its five aftershocks (for IG basin and Kathmandu). The ground motions are observed to get modified due to the presence of basin with peak ground velocity amplification of ~8-10 times due to the presence of the 3D basin. Though the method provides realistic ground motions and can be used to generate site-specific hazard curves, it is computationally expensive to run the simulations across the whole Indian landmass. In addition, the current seismic zone map recommended in the IS1893-2016 code is based on past damage intensities and common engineering judgments. However, because there is so much uncertainty about the location, size, and resulting shaking intensity of future earthquakes, a probabilistic seismic hazard map (PSHA) of the country is constantly needed. Hence, by including the best-known seismicity and source information, we develop a new seismic hazard map of India based on fault oriented elliptical smoothening approach in the traditional PSHA framework. Besides, the current IS1893-2016 design practice is to use vertical spectral acceleration as 2/3rd of the horizontal spectral acceleration in the design and analysis of structures. Here, we investigate the 2/3 exceedance of V/H for various periods and combinations of magnitudes and distances from hazard deaggregation. The study also attempts to arrive at a simplified bilinear approximation for displacement hazard response spectra, which is of prime importance in displacement-based design approaches. Apart from this, several parameters of engineering importance are explored, and their applicability in liquefaction and landslide hazard analysis forms an important basis for design requirements and risk analysis in the region