<p>This study presents an integrated approach for evaluating the performance of photovoltaic (PV) panels by combining nonlinear multivariate regression, artificial neural networks, and thermodynamic (energy and exergy) analyses under real-world operating conditions. Experimental data reveal that panel efficiency is most significantly influenced by maximum power output (<i>P</i>_max), solar irradiance, fill factor, and temperature, with strong nonlinear interactions among variables. Nonlinear regression models achieved high predictive accuracy (<i>R</i>² &gt; 0.98), outperforming traditional linear methods. Neural network analysis further confirmed the dominance of <i>P</i>_max and irradiance as key predictors of PV efficiency, demonstrating the capability of machine learning to capture complex dependencies in renewable energy systems. Importantly, the integration of exergy analysis added a thermodynamic perspective, showing that rising panel temperatures lead to increased irreversibilities and decreased exergy efficiency, even when energy input remains stable. This insight underscores the importance of temperature regulation techniques such as phase change materials and water cooling to enhance PV performance, particularly in warm climates. The nonlinear effects of humidity on efficiency also highlight the value of advanced modeling in understanding environmental influences often overlooked by conventional models. The findings contribute to the broader discourse on Energy and Climate Change by emphasizing the need for resilient, adaptive solar technologies capable of operating efficiently under increasingly variable climatic conditions. As the global energy transition accelerates, the integrative modeling framework proposed in this study offers a robust tool for optimizing PV systems to support low-carbon and climate-resilient energy infrastructure.</p>