The AlF3-MF2-LnF3 (M = Ca; Sr; Ba; Ln = Yb/Tm) based pure fluoride GCs are essentially ideal candidates for the realization of PL in UV and VIS spectral region. However, the pure fluoride glasses are prone to precipitation of undesired nano-crystals while quenching and post annealing due to their poor chemical stability. Therefore, the development of such glass systems requires strict adherence to composition control right from the outset. This research employed extensive molecular dynamics (MD) simulations and incorporated the results into the experimental design to develop an efficient multicomponent fluoride glass system. Vivid simulation of M-F-Ln and M-F-Al nano-cluster formation via MD are observed in the glass matrix. These nano-clusters serve as the structural origins of the precipitated nano-crystals in GCs post heat treatment. Through meticulous manipulation of the Al/M ratio, it became feasible to theoretically forecast the transformation of the glass matrix from a cubic M1-xLnxF2+x to a hexagonal Ln1-yMyF3-y nano-crystal precipitated GCs, which could conceivably be manufactured experimentally. This transformation offers superior luminescent performance for Ln1-yMyF3-y containing GCs enabling short-wavelength PL realisation in near UV (∼361 nm with 8-fold enhancement) and VIS (∼478 nm with 1.7-fold enhancement) from Yb3+/Tm3+ transitions, in comparison to its cubic M1-xLnxF2+x counterpart. Hence, we propose an effective fluoride glass system for short-wavelength upconversion luminescent applications through the integration of MD simulations and experimental research.