Safety management of hydrogen infrastructure is vital for sustainable progress in the hydrogen economy. Accordingly, this paper presents a dynamic and holistic risk model to address some significant shortcomings of the current hydrogen risk analysis models. The hydrogen release scenarios are modeled using the Bow-tie technique integrated with improved D Numbers Theory and Best-Worst Method. This helps to analyze epistemic uncertainty in the prior probabilities of the causation factors and barriers. Subsequently, a Dynamic Bayesian Network (DBN) model is developed to analyze dynamic risk and deal with aleatory uncertainty. The application of the proposed model is demonstrated on a water electrolysis process. The results of the case study provide a better understanding of the causal modeling of accident scenarios, associated evolving risks with uncertainty. The proposed model will serve as a useful tool for the operational safety management of the hydrogen infrastructure or other complex engineering systems. (C) 2020 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.