With the continuous increase in renewable energy penetration, the combined fluctuations from intermittent sources such as wind and solar power, along with load forecasting errors, have significantly heightened the uncertainty of grid net load. Traditional scheduling methods struggle to meet real-time balancing requirements under these conditions. To address this challenge, a real-time hydropower scheduling model for rapid response to net load deviations is proposed. The model features a two-layer objective function: the first layer minimizes the number of participating hydropower plants to streamline command transmission and execution, thereby reducing communication delays and coordination time and enhancing response speed; the second layer minimizes total water consumption by optimizing power output allocation among plants. A deviation consistency constraint is introduced to ensure hydropower output adjustments align with the net load deviation, preventing ineffective adjustments. The model is solved using mixed-integer linear programming. A case study in a southwestern basin demonstrates that the model reduces deviation balancing time by approximately 55%, decreases output adjustment frequency by about 24%, and lowers water consumption by nearly 23%, significantly enhancing the efficiency and economic performance of grid regulation.