In this thesis, structures and properties of hybrid perovskites are studied at the atomic scale, with a focus on their optoelectronic applications. Three-dimensional hybrid perovskites are mainly discussed in this work for solar photovoltaic cells and a two-dimensional hybrid perovskite for a photodetector in the last chapter of results. A solar cell made of methylammonium lead iodide (MAPI) or CH$_3$NH$_3$PbI$_3$ has previously been measured to have excellent power conversion efficiency which was attributed to the material's band gap. This understanding of the relationship allows us to tune the band gap by mixing compositions so the theoretical efficiency of the solar cells can be optimised. Here, we first investigate the structural, energetic and electronic properties of the potential candidates, MAPI, CH$_3$NH$_3$PbBr$_3$, CH(NH$_2$)$_2$PbI$_3$ and CH(NH$_2$)$_2$PbBr$_3$ (FAPB). The results suggest they can be present at various crystal phases, including narrow-band-gap cubic-like and wide-band-gap hexagonal phases. Yet, the hexagonal phase of these bulk perovskites is more favourable at room temperature so the higher band gaps limit the theoretical efficiency of the solar cells. %ranging from 1.55 to 1.85 eV. Two methods of mixing are studied to improve the properties for solar applications, such as solid solution and superlattice. In the solid solution, we take into account all studied 3D perovskites mentioned above as constituents and compared their hexagonal and cubic-like phases. The results show that while the intermixed cubic-like phases of all combinations remain to have band gaps within the range of those of the constituents, their structural stability improves, i.e. more energetically favourable than hexagonal structures at room temperature. Therefore, the solid solution is suggested to be used as an alternative to bulk perovskite to build the single-junction solar cell or used for the multi-junction solar cell to improve the measured efficiency. In the superlattice, MAPI and FAPB are considered in various stacking layers. Both cubic and hexagonal phases of the heterostructures are examined in terms of structural, energetic and electronic properties. Promisingly, the cubic structures have shown improvement in band gaps and FAPB(2)/MAPI(1) even exhibits the optimal value of 1.34 eV that predicts the maximum theoretical efficiency of a single-junction solar cell. Moreover, these superlattices undergo a phase transition from hexagonal to cubic at various temperatures below the room temperature and the energy required to form a cubic interface is within the thermal excitation energy at room temperature. These findings conclude that superlattice structuring of hybrid perovskites will outcome the most efficient single-junction solar cell. The layered two-dimensional perovskite studied as a photodetector is 4-fluoro-phenylethylammonium lead iodide. The bandgap is shown in good agreement with the experiment measurement by collaborators. We predict the Schottky barrier of the electron at the interface between the perovskite and Au contact. Several mechanisms are studied to explain the measured absorption peaks within the band gap and iodine vacancy in the perovskite shows a similar absorption peak. The extent of this work has demonstrated two methods of structuring the mixed hybrid perovskites to examine the band gap and structural stability for solar cell applications. Both studies have shown the stability improves but the investigation of superlattice also successfully has predicted a mean to achieve the optimal band gap. Moreover, we have also studied the properties of the two-dimensional hybrid perovskite to support its experimental measurements in a photodetector device.