In the physical modelling of coastal engineering problems, use is often made of foreshores and transition slopes to obtain the desired wave conditions - both spectral parameters and wave height distribution - at a given location. Numerical wave models can be used to predict whether the target wave conditions are met for a given physical model layout and wave forcing. The XBeach non-hydrostatic two-layer model is a computationally efficient numerical model that has been validated for spectral wave parameters, but lacks validation of the simulated wave height distributions. In this work, wave flume data with high spatial density over a sloping foreshore is used to validate the ability of this numerical model to reproduce both spectral wave parameters and wave height distributions. As part of this effort, optimal settings have been derived for the wave breaking formulation used in the numerical model, resulting in recommended values for the maxbrsteep and reformsteep parameters of 0.40 and 0.20 respectively.   From the results of the validation it is concluded that the numerical model is unsuccessful in reproducing the validation tests with 5.0% wave steepness, potentially due to the higher kph numbers on the generating model boundary. Hence, using the numerical model with values of kph ≥ 2 on the model boundary is not recommended.   The XBeach non-hydrostatic two-layer model performs much better for the 1.0% and 2.5% wave steepness tests, where the spectral wave parameters are represented well. The corresponding wave height distributions are represented reasonably well up to the point that the relative water depth gets very shallow. For shallower water, the model is expected to underestimate the higher waves in the wave height distribution. Additionally, the numerical model is shown to reproduce the wave height distribution better than a commonly used analytical formulation for wave height distributions on slopes.