Over the past years Fibre Reinforced Polymers, or FRP, have started making more of an appearance in civil engineering structures. They have the advantages that they are light-weight, do not corrode and theoretically require little maintenance during their lifespan. Originally they were used to construct reinforcement, cables and small bridges but more and more they are finding uses for larger scale structures. One of these newer applications is in lock gates. Locks are structures which are responsible for enabling water based transport while also retaining high water where necessary and are critical links in the water defence system of a region. Their gates also have a relatively large risk of collision due to the amount of moving vessels passing through them. In order to safely construct these lock gates from FRP laminates it is important that their response to such collision loads is well understood. This is the focus of this study with the aim to construct a model to help better understand the collision scenario. To make the theory concrete a case study is done on the lock gates of Sluis III which is situated in the Wilhelminakanaal in Tilburg. These gates are, at the time of writing, the largest FRP lock gates in the world. With the down stream gates being 13.9 by 6.3 meters. The event in which a Class III vessel collides with these gates will be examined in detail. The collision is schematised as a one dimensional collision using a series of springs and dampers to obtain understanding of the general collision behaviour. From this model it is concluded that the application of a load from the ship’s engine or taking elastic deformations of the ships bow into account has negligible influence of the results (in the order of 2%), simplifying the calculations considerably. This simple model is later advanced in three ways: Two numerical finite element models are used to determine the structural response of the gate elements on a global and local scale and a more advanced analytical model is made to account for non-elastic deformation in the ship’s bow. The gate’s structure consists of multiple overlapping laminates which together form the skins of the gate. The numerical model is set up in two ways, one of which is used to determine the overall response of the gate element and the other to focus on the interlaminar resin layer in the skins. The results show that it is of importance to apply the load in a realistic manner as the results may vary widely depending on bow shape and point of impact. The approximation suggested in the Dutch codes in which the load is applied as a distributed load of a 0.5m2 areas proved to be inaccurate. For this reason a dynamic LS-Dyna calculation is run using a rigid model of the ships bow to apply the load. The outcome of this analysis shows minor damage to the gate over a large area around to point of impact, but the stresses remain under the failure limit of the laminates except for the internal flanges directly under the impact. These flanges will fail, but this will not threaten the water retention of the gate. The stresses in the resin layer also remain under their critical limit. It can be concluded that the gate satisfies the requirements but significant repairs will be necessary to restore it to a fully operational state. The expanded analytical model is based on the fact that the force between the bow and the gate is larger than the failure load of the bow itself. The failure which will then take place will dissipate large portions of energy (in the order of 50%) making the current approach, in which this does not take place, highly conservative. The model suggested here is a segmented failure model in which parts of the ship bow fail completely once their failure load in reached. The results of this model are dependent on a series of inputs based on the ships structure and the damping during collision, but for all input values within their expected regions it shows a significant reduction in the amount of the energy that must be retained by the gate as well as a decreased sensitivity to the, hard to predict, damping factor. This model shows potential to reduce material usage for lock gates in which collision are considered governing. With further refinement this model could be used during the design of future lock gates to come to a cheaper design. Experimental data would serve an important purpose during this refinement.