Abstract: Coordinated operation of sluice gates in tidal river networks is a crucial measure to enhance hydrodynamic performance, reduce oscillatory flows, and improve water quality. However, current sluice operations primarily rely on empirical approaches, which tend to be inefficient and time-consuming when deciding on the optimal operation strategy. Existing intelligent approaches are mostly restricted to deterministic scenarios and lack the capability for flexible regulation of hydrodynamic processes under the combined influences of hydrology, tides, and gate-induced disturbances. In this study, a real-time decision-making optimization method is proposed for operational strategies in hydrodynamic regulation through sluice operations. Incorporating Long Short-Term Memory (LSTM) networks and Temporal Fusion Transformers (TFT), the method accurately captures long-term dependencies and non-stationary dynamic characteristics of water level and discharge variations, identify multi-source disturbances, and establishes nonlinear coupling relationships governing water level and discharge responses to enable precise forecasting. Additionally, a Genetic Algorithm (GA)-based optimization model for sluice gate group regulation is developed to maximize the predicted net water volume through a target cross-section by searching for the optimal gate operation scheme based on the water level difference between the upstream and downstream sides of the gates. The integrated model is applied to the hydrodynamic regulation of the Wennan watershed river network in Shanghai and performs well in dynamically responding to variations in hydrological and tidal boundaries. It intelligently provides a sluice gate group control scheme for the hydrodynamic reconstruction of the river network. Under two representative scenarios, the 12-hour net water exchange increased by 12.8% and 5.4%, while oscillatory flows were reduced by 74.2% and 55.3%, respectively, thereby significantly reduces oscillatory flow in the target reach, and enhances directional flow.