Mangrove forests cover large parts of the tropical and subtropical shorelines and grow in the inter-tidal zone. Mangrove forests act as a natural coastal protection. The most important features of this protection are the prevention of erosion by the collecting of sediment and the reduction of the wave climate. There is a need to quantify and model hydraulic and morphologic processes, wave attenuation in particular, in order to predict and simulate the natural hydraulic behaviour and evaluate the consequences of any artificial measure in mangrove forests. In this study it is found, that the best numerical approximation to model wave attenuation in vegetation with SWAN (spectral wave model), was a formulation described by Dalrymple (1984). The formulation expresses the amount of dissipated wave energy in terms of vegetation characteristics and wave parameters. It has been implemented in SWAN, resulting in an adapted SWAN model in which wave attenuation can be modelled due to homogenous, horizontally variable or vertically variable vegetation. Subsequently the implementation for homogenous vegetation was validated on the results of physical model tests from Mendez and Losada (2004). This validation was successful and gave evidence the implementation was done correctly. The drag coefficient (Cd) is the only calibration parameter in the Dalrymple (1984) formulation and depends on the vegetation characteristics and wave parameters. For the vegetation characteristics it turned out that the drag coefficient depends mainly on the relative spacing. The dependency of the drag coefficient on the wave parameters is determined by the horizontal orbital velocity and the peak period, and could be best described by the relation between the drag coefficient and the Keulegan-Carpenter number. Next a sensitivity analysis was performed which resulted in the expected relations between the wave attenuation and the vegetation characteristics, respectively the hydraulic conditions. In a test simulation with patchy vegetation it was found to be possible to model wave induced setup causes by the energy dissipation in the vegetation, however the potential flow itself can not be modelled by SWAN, because the model cannot simulate flow. Two measurement campaigns in Vietnam on wave attenuation in mangrove forests were modelled as test cases in the adapted SWAN model. In these cases hydraulic conditions and vegetation characteristics were known from measurements and thus the drag coefficient could be calibrated. Finally it is concluded that the implementation in SWAN was successful for different types of vegetation. Nevertheless the dependency of the drag coefficient on both wave parameters and vegetation characteristics implies that the model can only be used when vegetation characteristics and wave parameters have been measured and related to the drag coefficient.