This paper investigates a water level control model for the Great Lakes based on genetic algorithm optimisation. The Great Lakes and their surrounding water systems constitute a huge watershed covering many large cities, and it is crucial to manage the inflow and outflow of lake water. In this paper, a mathematical model of lake water level and flow is gradually established through a network flow model, taking the Great Lakes and their connected rivers as nodes and edges, and integrating the effects of multiple factors on lake water level, such as temperature, wind, tide, precipitation, evapotranspiration, river flow, water storage reservoir policy, seasonal cycle and longterm climate change. The genetic algorithm was used to balance the diversity needs of different stakeholders, and the optimisation objective function was established and the constraints were comprehensively set. The model results show that the algorithm can determine the optimal water level of each lake in different seasons, and reasonably satisfy the requirements of stakeholders such as residents, shipping, environmental protection, agriculture and fishery. The model provides a scientific and dynamic lake water management strategy, which provides a theoretical basis for coping with the complex and changing environmental conditions and multi-stakeholder conflicts.