This paper proposes a novel model-free control (MFC) strategy for hybrid electric vehicles (EVs) powered by a proton exchange membrane fuel cell (PEMFC) and a supercapacitor (SC). Unlike conventional model-based approaches that depend on accurate system identification and parameter tuning, the proposed framework employs ultra-local models to adapt dynamically to system variations without explicit modeling. The hybrid architecture is implemented using an interleaved boost converter for the PEMFC and a bidirectional buck–boost converter for the SC, coordinated to supply propulsion power and enable regenerative braking. Comprehensive MATLAB/Simulink simulations demonstrate that the proposed MFC achieves <3% current tracking error for both PEMFC and SC, ~750 ms settling time for PMSM speed variations, and <120 ms response for power transitions, while the DC bus voltage remains tightly regulated under dynamic load disturbances. Hardware-in-the-loop (HIL) validation on an OPAL-RT 5600 platform further confirms the method’s feasibility, showing a 20% reduction in execution time and enhanced robustness against parameter uncertainties compared to classical PI control. Experimental results also verify stable current sharing in interleaved converters, accurate voltage regulation in the SC branch, and smooth torque generation in the PMSM drive. Overall, the proposed control strategy provides a computationally efficient, fault-tolerant, and plug-and-play solution for next-generation EVs by reducing calibration effort and ensuring reliable operation under nonlinear and uncertain conditions, while demonstrating clear potential for real-time automotive applications.