Tsunamis are a series of long waves, caused by rapid perturbations of the water level. For Banda Aceh, only submarine earthquakes are a likely cause of tsunami-generation and only negative waves (with an initial decrease of the water level) will arrive at the shore line. The probability of tsunamigenic earthquakes in Banda Aceh was investigated, based on existing models. The propagation of a tsunami wave can be described with equations that are valid for normal shallow water waves. Difficulties arise in determining the initial signal (excitation) in relation with earthquake magnitude. Furthermore, simple shoaling laws are not accurate close to the shore. Therefore a depth-averaged 2D-model is used to determine the wave heights at the Banda Aceh shoreline for different earthquakes magnitudes (see Table above). The reference case is the Dec2004 Tsunami that caused almost 300.000 missing or dead people in the countries around the Indian Ocean. In Banda Aceh alone, more than 70.000 people were reported dead or missing and the damage was estimated to 1,12 billion USD. It is concluded that for higher tsunamis, low-crested structures and protection by mangrove trees or other vegetation is not effective. Protection should aim at reflection of the tsunami wave. Therefore, three main tsunami protection alternatives are developed; 1) a tsunami barrier located offshore at 10m water depth, 2) a coastal barrier and 3) an inland barrier. These structures are designed and modelled in such a way that overtopping could occur. Many sub-alternatives are modelled, both in numerical (depth-averaged) 2D and 1D-models. The effectiveness of each alternative is expressed in the amount of inundation volume that still entered the area of Banda Aceh compared to the Dec2004 Tsunami event. It was found that offshore structures show the highest reduction of the tsunami wave in relation to their retaining height (i.e. the visible height of the structure). A procedure is set up to relate tsunami inundation volume with damage. With decreasing protection level the residual damage increases. By relating this residual damage with the probability of tsunami-events, the residual risk for each protection level is defined. Based on typical designs for the three alternatives, construction costs are estimated. By adding up the (Net Present Value) of the residual risk and the construction costs, the optimal design level for the three alternatives is derived. For these optimal levels, the costs of tsunami protection are compared with the benefits (i.e. the prevented risk). From this Cost Benefit Analysis it is concluded that none of the alternatives is economically feasible, except under very favourable conditions. In this case only the inland and coastal alternatives show a positive cost-benefit ratio. For the offshore barrier, the costs consequently outweighed the benefits. Nonetheless, considerations about the 1) little escape time between a tsunami warning and the actual arrival, 2) the value of life, and 3) other non-quantifiable advantages of a tsunami protection structure, could still support a decision in favour of an extensive tsunami protection. In that case, a Multi Criteria Analysis showed that offshore solutions are most preferable. High structures on land will unacceptably affect the socio-economic development of Banda Aceh. Thereby, tsunami height, run-up, velocities and impact are smaller in offshore conditions. Therefore it was decided to elaborate further on the design of an offshore tsunami barrier, despite the negative cost-benefit ratio. Finally a design is presented that is stable in earthquake conditions and can withstand the high current velocities both over the barrier (overtopping) and through the gaps.