Every year, there are many floods on the planet, which have a significant impact on ensuring the safety of people and affects the quality of life. The development of modern modelling technologies makes it possible to predict various scenarios for the development of the situation and reduce the likelihood of negative consequences. This issue is especially relevant for settlements located in the immediate vicinity of hydroelectric power plants, since by regulating discharge costs from hydroelectric power plants, it is possible to safely pass flood waters avoiding flooding of residential buildings and infrastructure, but this requires knowing the flooding zones at different water levels and discharge costs. This paper presents the results of solving the problem of modelling the dynamics of flood waters within the boundaries of the settlement of Krasnoyarsk. To calculate the flooded areas, the TUFLOW program was used in the Surface-water Modelling System modelling environment, as well as neural network forecasting using the NeuroPro software product. The simulation results made it possible to predict local flooding of the settlement during the flood of 2021 and take preventive measures to reduce the risk of flooding.