Currently, the energy utilization rate of wind-solar-hydrogen energy system is low, and the traditional capacity optimization is modeled based on the fixed hydrogen production efficiency of the electrolyzer and only takes into account the existence of one operator in the system, without considering the competition and cooperation between different operators, the change of hydrogen production efficiency during the operation of the electrolyzer, and the standby start-up cost, which leads to unsatisfactory capacity optimization results. In order to solve the above problems, the organic Rankine cycle is firstly combined with the fuel cell for waste heat recovery; a mathematical model of dynamic hydrogen production efficiency is constructed by analyzing the influencing factors of hydrogen production efficiency of the electrolyzer, and binary variables are introduced for describing the standby startup cost of the electrolyzer; a capacity optimization method based on game theory is adopted, and the Nash equilibrium of the game is solved by using the sparrow search algorithm with the objective of maximizing the annual revenue. The simulation results show that: the total revenue of the energy system increases by 4% after adding waste heat utilization; the capacity of the proton exchange membrane electrolyzer after considering the dynamic hydrogen production efficiency and standby startup cost decreases by 26.1%; the revenue of the hydrogen storage power station decreases by 55.3%; and the coalition is the most stable with the largest revenue and the most reasonable capacity allocation when the three parties, namely, wind solar and hydrogen, form a grand coalition.