The digital twin technology has been developing rapidly in the water resources field. This paper outlines how hydrodynamic models underpin the digital twin of water resources, and proposes the requirements for advancing hydrodynamic models and suggests corresponding model solutions. Key techniques for enhancing the capabilities of typical models in digital twins of water resources, such as 1D free surface and/or pressurized hydrodynamic models, are introduced. Techniques for improving the computational efficiency of 2D hydrodynamic models in large areas with high resolution are also described. Specifically, the mathematical challenges, key techniques and applications of recently developed models inclusive of coupled hydrodynamics and sediment transport models for the breaching processes of dams and dikes as well as hydrodynamic models for flash floods aiming at optimizing early warning systems are expounded. Finally, directions for further development of hydrodynamic models are proposed, including the need for higher-dimensional models and coupled models involving different dimensions, models for multi-phase processes, further investigation into evaluation and real-time modification techniques for hydrodynamic modeling, development of integrated models combining physically-based and data-driven approaches, and the establishment of mutual connections and feedback models for virtual modeling and physical reality control.