This paper offers a state-of-the-art energy-management strategy specifically developed for FCHEV focusing on robustness under uncertain operations. Currently, energy management strategies try to optimize fuel economy and take into account the sluggish response of fuel cells (FCs); however, they mostly do so assuming all system variables are explicit and deterministic. In real-world operations, however, a variety of sources may cause the uncertainty in power generation, energy conversion, and demand interactions, e.g., the variation of environmental variables, estimated error, and approximation error of system model, etc., which accumulates and adversely impacts the vehicle performance. Disregarding these uncertainities can result in overestimation of operating costs, overall efficiency and overstepped performance limitations, and, in serious cases can cause catastrophic system breakdown. To mitigate these risks, the current work introduces a neural network-based energy management system (EMS) that guarantees both optimal performance and resilience when confronting uncertain operating conditions. In contrast to classical model-based control techniques that can falter when faced with the unpredictable nature of real-world environments, the proposed NN-based EMS continuously recalibrates in response to evolving variables, thereby enhancing the reliability of performance and the practicality of control actions.. The new approach embeds the uncertainty directly into both the cost function and the constraints, significantly reducing the risks tied to making less-than-optimal choices. By also introducing adjustable conservatism level parameters, it offers a way to fine-tune the trade-off between pursuing the best performance and ensuring robustness, which makes the energy management strategy more responsive and flexible. Utilizing the learning and generalization strengths of neural networks, the method improves the day-to-day practicality of energy management in fuel-cell hybrid electric vehicles, delivering operation that is not only more efficient and stable but also more cost-effective in real-world deployments.