This study uses an experimental and machine learning-based approach to investigate the effects of hexagonal honeycomb structures on forced convection heat transfer. The research addresses the lack of comprehensive studies on hexagonal geometries' thermal and hydraulic performance under varying flow and current conditions. This study's novelty lies in integrating experimental and machine learning methods to optimize heat transfer and friction characteristics, contributing to the efficient design of heat exchangers for industrial applications. An experimental setup with a turbulent flow regime and aluminum honeycomb heat exchangers was employed. The system parameters, including Nusselt number (Nu) and friction factor (f), were analyzed for various honeycomb geometry rates (0.26, 0.56, and 0.80), airflow velocities (10, 15, and 20 m/s), and electrical currents (50, 100, 150, 200, and 250 A). The data were processed using Artificial Neural Networks (ANNs) with a scaled conjugate gradient (SCG) algorithm, achieving high accuracy in performance indicators such as MAE, MSE, ARD%, and R2. Experimental results showed significant improvements in Nu with increasing current and flow rate, while the impact on f was minimal. The highest thermo-hydraulic performance criteria (PEC) was observed for HGR = 0.80 at 20 m/s and 250 A, with a Nu increase of 36.46 % compared to the baseline. As a result, the findings demonstrate the potential of hexagonal honeycomb structures to enhance thermal performance while maintaining low-pressure drops, making them suitable for high-efficiency energy systems.