<p>In this study, a novel approach was developed by integrating numerical techniques with machine learning algorithms for distributed flood routing. This combination of numerical simulations and machine learning provides a pioneering method for predicting flood routing, significantly improving both accuracy and reliability. Using OpenFOAM, the nonlinear Navier-Stokes partial differential equations were solved to precisely model flood routing behavior. Additionally, artificial neural networks (ANNs) trained on data from OpenFOAM simulations were employed to successfully forecast flood routing outcomes. Sensitivity analyses were conducted on both hydrological and geometrical parameters to assess the effectiveness of the OpenFOAM model. The results indicate that as the slope increases, the relative change in peak flow decreases. Similarly, as foundation depth increases, the variations in peak flow along the canal diminish. Moreover, when the riverbed depth is shallow, the relative reduction in maximum depth is more pronounced than when the riverbed depth is greater. The study concluded that ANNs with ten neurons in the hidden layer demonstrated optimal performance for flood routing analysis. While this research was carried out in a laboratory-simulated channel, future work should include dimensional analysis and real-world river studies to enhance its practical applicability.</p>