Wave climate in harbours determines for a significant part the efficiency of harbour operations. A good prediction of the hydrodynamics is therefore regarded as indispensable, which is often done by means of advanced numerical modelling. This study is a part of the Joint Industry Project XBeach. The goal of this research is to determine the application limits of the numerical model XBeach (both nonhydrostatic and surfbeat) for the purpose of modelling hydrodynamics in harbours. XBeach non-hydrostatic, a phase-resolving model where the waves are resolved on the individual wave scale, is theoretically able to predict the hydrodynamics because relevant short and longwave processes are resolved. To this end, the depth-averaged nonlinear shallow water equations are solved. Further, the application range for phase-averaged models is smaller, because the waves are averaged and no phase information is considered to include the dominant wave processes in harbours (e.g. diffraction). XBeach surfbeat (Roelvink et al., 2009) is a mix between phase-resolvingmodel (long waves) and phaseaveraged model (short waves). This indicates that the long wave processes (e.g. harbour oscillations) are included in surfbeat. This mode uses the energy balance equation, which is coupled to the nonlinear shallow water equations for the long waves. To be able to examine the applicability of XBeach, two simplified harbour layouts are considered which are subject to a variety of wave conditions. The considered governing wave processes are reflection, diffraction and harbour oscillations. These laboratory experiments are reconstructed in XBeach and its results are compared to the measurements. Wave-structure interaction is not further considered, because the XBeach model does not have the possibility to define structure properties. Comparison between the simulation results and measurements have shown that periods and wave heights of the short waves (f > 0.04Hz) are only well calculated by XBeach non-hydrostatic. Wave conditions with a relatively large wave height over depth ratios (up to 0.3) are still accurate. XBeach surfbeat gave inaccurate results for the wave heights of the short waves, because diffraction is not considered in the calculation model. With respect to the long waves (f · 0.04 Hz), both modes in XBeach are able to reproduce the wave heights, wave periods and the resonant frequencies. The accuracy of the phase-resolving and phaseaveraged modes is comparable. However, more confidence is placed in XBeach non-hydrostatic, especially when long-crested waves are involved. The accuracy of XBeach surfbeat for the wave heights is higher if the long waves are still bounded to the wave groups inside the harbour. Without the presence of these short waves, the long waves are free which can freely amplify resulting in an overestimation.