Perovskite solar cells have emerged as the most promising cheaper photovoltaic technology. Besides solar cells, halide perovskites have a wide range of applications due to their remarkable optoelectronic properties. Starting from 3.8% in 2009, solar to the power conversion efficiency of perovskite solar cells is now >25%, exceeding market leader polycrystalline silicon. These advancements are mainly due to the improvement in their manufacturing process targeted to make better morphology. To this objective, tens of thousands of experimental studies are conducted where a wide range of additives, compositions, and processing conditions are optimized based on a trial and error methodology. Although this approach increased the efficiency, however, perovskite electronics suffer from the problem of stability and reproducibility. Therefore, to industrialize highly efficient perovskite electronics, one needs to understand better and rationalize their fabrication process, i.e., crystallization of halide perovskites. In this thesis, I perform molecular simulations to understand the formation process of halide perovskites. First, I try to lay out the atomic details of the nucleation process of perovskite from the solution. I find that nucleation proceeds in a two-step manner where an intermediate phase forms before the emergence of perovskite crystals from the solution. I observe that monovalent cations can play an essential role in the nucleation process. On this information, simulation-driven experiments are performed by experimental collaborators to demonstrate that the grain size and crystallinity can be tuned by only varying the concentration of monovalent cations in solution. I also attempted to study the effects of different solvents and pseudo-halide additives used to make high-efficiency solar cells. I calculated that by modulating the solute-solvent interactions, one could minimize raw material losses during the coating process. In comparison, pseudo-halide additives are found to passivate the ionic defects and may slow down the growth process that helps to improve the final morphology. Second, I study the solid-solid phase transition in the popular two-step process. I discover a new path to the low-temperature formation of highly efficient metastable cubic phase of formamidinium lead iodide perovskite. I first validate my simulations with in-situ experiments performed by experimental collaborators. Then simulations are employed to design a practical methodology to successfully make thin films of phase pure cubic formamidinium lead iodide at lower temperatures. Moreover, I study the intrinsic stability of this material. I calculate the energy barrier between the metastable photo-active phase and its photo-inactive thermodynamically stable hexagonal phase. I find that once formed; the photo-active phase might be kinetically stable. Additionally, I identify new polymorphs (perovskitoids) of halide perovskites and also study the structural insights of low-dimensional halide perovskites. Overall, I present a fundamental understanding of the crystallization of perovskites which is helpful to design better experiments to make high efficiency and stable perovskite electronics.