Hot water-based thermal energy storage (TES) tanks are extensively used in heating applications to provide operational flexibility. Simple yet effective one-dimensional (1-D) tank models are desirable to simulate and design efficient energy management systems. However, the standard multi-node modelling approach struggles to reproduce the dynamics of highly thermally stratified tanks due to their artificial numerical diffusion. In this paper, a novel 1-D multi-node modelling approach is introduced for accurately simulating water tanks with a high extent of thermal stratification. A non-linear, hybrid continuous–discrete time model able to capture the sudden temperature change within the tank is presented. The modelling approach was adopted to simulate a commercial TES tank, with the model being implemented in MATLAB/Simulink. Results from experimental tests were compared with simulation results, demonstrating that a hybrid continuous–discrete 12-node model accurately estimates the temperatures of the tank. It is also shown that the hybrid model avoids the numerical diffusion exhibited by standard multi-node models. This has been evidenced by the reduced root mean square and mean absolute errors exhibited by the hybrid model when compared with the experimental data.