<p>This paper presents a deep learning–based hybrid energy management strategy for a Permanent Magnet Synchronous Motor (PMSM) drive powered by solar, battery, and supercapacitor. The proposed system aims to improve both energy efficiency and speed tracking performance under varying driving conditions. In the hybrid system, battery supplies continuous long-term power, while the supercapacitor supports sudden transient demands during acceleration and rapid load variations. Solar PV further reduces the dependence on stored energy. A custom neural network controller is employed to coordinate power distribution and regulate the nonlinear dynamics of the PMSM drive based electric vehicle. Comparative analysis with neural network fitting and neural network time-series controllers demonstrates that the proposed controller achieves faster convergence, lower mean squared error, and improved robustness under dynamic operating scenarios. The training analysis further shows that controller performance is strongly influenced by the selection of learning rate and training epochs, where suitable hyperparameter tuning significantly improves accuracy and convergence behaviour. The PMSM model is parameterized using the Finite Element Method to provide a more realistic representation of the nonlinear behaviour as magnetic saturation, <i>dq</i>-axis cross-coupling, and parameter variations. This improves the responsiveness and reliability of the motor under practical operating conditions. The effectiveness of the proposed hybrid energy management strategy is validated through MATLAB/Simulink-based analysis, which confirms stable power sharing, reduced battery stress, and efficient system operation. The proposed system is further evaluated under standard and regional driving cycles including NEDC, UDDS, WLTP, MIDC, Delhi, and Pune cycles, demonstrating consistent tracking performance and adaptability across diverse driving environments.</p>