Hydrogen leakage is a crucial risk for the hydrogen generation unit, which would lead to the potential fire and explosion accidents. Hydrogen leakage risk analysis is the essential alternative to ensure the safety of the hydrogen generation process. This paper presents a dynamic risk analysis methodology regarding the hydrogen leakage in the hydrogen generation unit by using the dynamic Bayesian network, which is employed to address the potential uncertainty and dynamic nature underlying the leakage risk of the hydrogen generation unit. A case study of hydrogen generation unit is carried out to demonstrate the applicability and advantage of the proposed methodology. Results indicate that the leakage probability of hydrogen generation unit can be significantly decreased within one year through equipment repair. Furthermore, the failure and repair rates of overflow alarm and pressure sensor are the most contributory factors to the hydrogen generation unit leakage. Finally, some active mitigative suggestions are presented to further reduce the leakage risk of the hydrogen generation unit. (C) 2019 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.