<p>Physics-Informed Neural Networks (PINNs) offer a new dimension to solve problems related to water wave dynamics by allowing the incorporation of sensor data. In this study, the application of PINNs is extended to model deep-water linear wave propagation where wave generation and absorption are implemented in the model at the inlet and outlet boundaries, respectively. The governing equations, boundary conditions and sensor data are integrated into the loss function that is minimized during the training process. The Volume of Fluid (VoF) method is employed to capture the air–water interface, representing the surface elevation of the propagating waves. The model predictions are validated against a Computational Fluid Dynamics (CFD) simulation developed in OpenFOAM. A hyperparameter study and hyperparameter tuning were conducted to study the relationship between each hyperparameter and the model’s performance. Through this process, the number of neurons, activation functions and loss component weighting are identified as the key hyperparameters that had the largest influence on model performance. We show that with optimized hyperparameters, the model is capable of accurately modelling the wave propagation from generation to absorption. The extension of PINN architecture to a more general and higher fidelity computational model, as described in this work, paves the way for PINN to be applied to more complicated and practical problems.</p>