Metal Inert Gas (MIG) welding is pivotal in industrial applications due to its efficiency and quality,
with solidus temperature critically influencing weld integrity. This study investigates the effects of
current (240–270 A), voltage (23–26 V), and wire feed rate (2.4–3.0 mm/s) on solidus temperature
of mild steel metal plate, integrating experimental data, Response Surface Methodology (RSM), and
an Artificial Neural Network (ANN) to develop predictive models. Experimental results revealed
that current and voltage significantly affect solidus temperature, peaking at 1642°C (260 A, 25 V,
2.8 mm/s). ANOVA validated the quadratic RSM model’s robustness (R²=0.9988), while the ANN
model further reduced prediction errors to <1°C. Contour plots elucidated parameter-temperature
relationships, demonstrating synergistic interactions. These models enhance process optimization,
offering precise control over weld quality. The ANN’s superior accuracy highlights machine
learning’s potential in advancing welding parameter prediction. Findings provide actionable
guidelines for achieving consistent, high-quality welds in industrial settings. Future work may
expand parameter ranges and integrate advanced machine learning techniques to refine predictive
capabilities, fostering smarter, adaptive welding systems.