In laser spot welding, FEM-based simulations often yield inaccurate temperature distributions, posing challenges in predicting leftover stress and grain traits. This study addresses these challenges by introducing a novel methodology that rigorously solves transient conservation equations governing mass, momentum, and energy. Key boundary conditions, such as the Marangoni effect and no-slip conditions, were incorporated to improve accuracy. Numerical findings demonstrate that the size of the laser beam significantly impacts the temperature distribution, with larger laser beam achieving steady-state conditions faster. Additionally, this approach accurately predicts stress distribution and grain characteristics by capturing transient thermal behaviour and the effects of solidification kinetics under non-equilibrium conditions, including undercooling and recalescence. The results underscore the methodology’s potential for improving precision in welding simulations, providing insights into optimizing heat source parameters for better structural and thermal outcomes.