Electric mobility systems demand fast output voltage regulation for reliable motor performance and source protection, while also requiring long-term objectives such as source endurance. Hybrid fuel cell (FC)-battery systems are commonly adopted, and their coordination often relies on compact multi-input converters. However, conventional PI controllers, though effective for average regulation, require complex modeling in coupled structures and cannot strictly enforce voltage and current limits. Model predictive control (MPC) addresses multiple objectives but faces a tradeoff: rich cost functions are too computationally heavy for high-frequency execution, while simplified short-horizon MPC cannot capture long-timescale objectives. This work proposes a multi-layer MPC for an integrated FC-battery converter. A fast one-step MPC regulates voltage and enforces source limits at the switching rate, while a slower supervisory layer improves endurance by updating reference values, such as moderating FC current rates. Effectiveness is validated via UAV simulations with a 450 W fuel cell, a 12-cell Li-Po battery, and a 16.8 V 400 W motor load.
