DC motor speed control is fundamental in modern industrial and robotic systems, where high precision, robustness and energy efficiency are required. Conventional integer-order Proportional–Integral–Derivative (PID) controllers often fail to capture the non- linearities and parameter variations inherent in real DC motors. This study proposes a control framework combining fractional- order (FO) system identification with an optimised fractional-order proportional-integral-derivative (FOPID) controller. The five FOPID parameters are optimised using four metaheuristic algorithms: Grey Wolf Optimizer (GWO), Firefly Algorithm (FA), artificial bee colony (ABC) and ant colony optimisation (ACO). Experimental validation on a MATLAB/Simulink R2024 (The MathWorks, Inc., Natick, MA, USA), an Arduino board, and a DC motor platform demonstrates that the particle swarm optimisation (PSO) FOPID controller achieves a settling time of 1.08 s with 2.00% overshoot and a control effort of 1.8 V/√s. Compared to the extended Ziegler– Nichols tuned FOPID, the PSO approach achieves 98.57% faster settling while maintaining comparable overshoot and demonstrating superior energy efficiency. Among the metaheuristic algorithms tested, PSO demonstrates the best overall performance with the lowest identification error and the most energy-efficient control effort. These results confirm the superiority of the metaheuristic optimisation approach over conventional tuning methods in terms of dynamic response, precision, and robustness for fractional- order control systems.