The popularity and adoption of phosphor-converted light-emitting diodes (pc-LEDs) have been enabled by the invention of GaN-based blue LED. Subsequently, using a blue or near-ultraviolet LED combined with a phosphor to produce white light has dramatically increased over the years allowing sizeable worldwide energy savings and leading to numerous LED phosphor intellectual property (IP). These phosphor-related IPs have since been the primary force driving phosphor innovations. A significant part of these innovations revolves around discovering novel phosphor compositions with competitive quantum efficiency, high thermal stability, and a practical photoluminescence spectrum. The conventional approach to develop a phosphor in experiment relies on searching for suitable host crystal structures amongst the well-known inorganic materials present in databases such as the Inorganic Crystal Structure Database. While thousands of materials have already been examined over the last two decades, only a few compositions have resulted in competitive and practical phosphor candidates. Alternatively, high-throughput screening based on first-principle calculations have proven successful in numerous disciplines. However, our current understanding of structural, chemical, and electronic factors that dictate luminescence properties, such as the emission bandwidth, wavelength, quantum efficiency, and thermal quenching remain qualitative and hinders the rational design of phosphors. Hence, this thesis addresses two essential characteristics of inorganic phosphors: the thermal quenching mechanism and the emission process of inorganic phosphors. In the first project, using first-principle calculations and ab-initio molecular dynamics simulations, two prevailing theories, the crossover, and thermal ionization mechanisms, were unified into a single predictive model for the thermal quenching of inorganic phosphors. The local environment stability of the activator was demonstrated for the first time to be related to the TQ under the cross mechanism. Furthermore, starting from the point charge electrostatic model, we account for the effect of the crystal field interactions on the thermal ionization barrier and show that a unified model can predict the experimental TQ in 29 known phosphors within a root-mean-square error of 3.1-7.6%. Finally, we develop an efficient topological descriptor derived from the activator's local Voronoi grid representation to rapidly assess the activator's local environment rigidity; hence, enabling a fast screening of vast chemical spaces to discover novel and thermally robust phosphors. In the second project, we have carefully reviewed the current theory of the Eu2+/Ce3+ photoluminescence mechanism and identified its current limitations and transferability to broader chemical spaces. Subsequently, the legacy covalo-electrostatic interaction model was reformulated into a harmonic parameterization of the emission energy where we uncover fundamental structural, chemical, electronic descriptors. As a result, the covalo-electrostatic harmonic parameterization (HCEP) model predicts the emission wavelength of Eu2+-activated nitrides with excellent accuracy of 7 nm. To further demonstrate the unprecedented capability of the HCEP-model, we conduct a multidisciplinary strategy consisting of an emission-targeted high-throughput screening via the HCEP-model, advanced experimental validations, and HCEP-guided site-engineering manipulations. More importantly, this screening was explicitly targeted to the near-infrared region, whereas only three poor efficiencies (IQE