Self-alignment is a critical phenomenon in Surface Mount Technology assembly, enabling precise component placement during solder reflow. This process, driven by surface tension forces, aligns components to minimize misalignment, which is essential for achieving high-quality manufacturing outcomes. Surface Evolver, a commonly used tool for studying self-alignment, models the behavior of liquid interfaces based on energy minimization principles. However, its application is constrained by model complexity and computational intensity. To address these challenges, this study introduces a simplified Energy Minimization Logic Simulator for analyzing self-alignment in a two-dimensional framework. The model simulates solder paste movement toward copper pads and component shifts toward solder paste using a gradient-based approach grounded in energy minimization principles. Using simulation data, the model learns the underlying physics of component movement under varying conditions of solder paste misalignment and volume. It then gradually refines its predictions by tuning parameters based on experimental data. The model achieves RMSE of 6 and 8 in the length and width directions, respectively. This combined approach of physics-based modeling and machine learning offers a novel method for optimizing chip placement. This study highlights the potential of integrating simplified 2D physical simulations with data-driven learning frameworks to enhance SMT assembly processes.