To address the severe load interruption and economic losses caused by extreme disasters in port power distribution networks, this paper proposes a post-disaster emergency dispatch strategy for hydrogen-powered port distribution networks considering frequency constraints. First, an operational model for port hydrogen energy systems is established through quantitative analysis of reversible solid oxide cells and hydrogen storage stations, along with their interdependencies. Building upon this foundation, a multi-timescale emergency dispatch methodology is developed by incorporating electricity-hydrogen supply-demand relationships. Specifically, for long-term scheduling, nodal loads are subdivided into power-continuously adjustable controllable loads and uncontrollable loads determined by feeder switches, with multi-type load interruption constraints formulated. For short-term scheduling, frequency stability constraints are established considering the stochasticity and volatility of renewable energy generation and loads. Subsequently, the big-M method is employed to transform nonlinear constraints into a classical mixed-integer linear programming problem for solution. Case studies are conducted on a modified IEEE 33-node distribution network. Simulation results demonstrate that the proposed strategy can effectively reduce post-disaster losses while ensuring power supply reliability and grid security, offering a novel perspective for post-catastrophe energy restoration.