Abstract: To address the problems such as limited regulation resources, high regulation costs, and slow response speed in traditional distribution network voltage control methods, a two-stage stochastic optimal control strategy with participations of power-to-hydrogen (P2H) devices is investigated. Firstly, voltage regulation device operation constraints and distribution network line constraints are modeled. Then a two-stage day-ahead and intra-day voltage optimal control model is developed considering electrolytic gas production revenue. Secondly, to deal with the voltage fluctuations or even over-limitation problem caused by short-term disturbances of renewables and load demands, typical operation scenarios of distribution networks are constructed by adopting Latin hypercube sampling and Kantorovich distance reduction techniques. Then the voltage control strategy is solved by minimizing the expectation of intraday-stage objective functions under all scenarios. Finally, case studies have shown that compared with the traditional voltage methods without considering P2H participations, the voltage over-limitation can be effectively avoided and the total regulation costs can be reduced by over 26.43% by using the proposed method.