<p>This research explores the design of emergency supply chains across risk-level categorised regions, addressing differentiated flow restrictions—a topic understudied in traditional supply chain literature. A mathematical model for network optimisation is developed, integrating risk classifications and dynamic flow constraints. Formulated as a bi-objective model, it considers both time and cost factors. The model employs stochastic mixed integer linear programming to handle random parameters (e.g., time distributions) and conducts sensitivity analysis on sub-objective weights to examine optimal solutions under varying priority settings. Empirical results show that, regardless of weight configurations, the supply volume within same-risk regions exceeds 80%, indicating that design should prioritise intra-risk-level regional supply while allowing necessary cross-risk-level flows. This work contributes to the theoretical framework of emergency supply chain design and provides practical guidance for regional industries to coordinate emergency responses, demonstrating its value in balancing efficiency and resilience under risk-based restrictions.</p>