This paper shows that model predictive control (MPC) in combination with moving horizon estimation (MHE) can more effectively be used for flood prevention in comparison with a three-position controller (TPC). Because MPC takes rain predictions into account, it uses the buffer capacity of the available flood basins in a more optimal way. Simulation results for historical data for the Demer, a river in Belgium with a history of large floodings, show a significant reduction of the number and the impact of floodings when MPC is used. Because of the strong nonlinear river dynamics it is necessary to use nonlinear MPC. The use of a state estimator like MHE is necessary because only some of the states are measured.