In this paper, structural optimization of a floating breakwater is addressed through several methods concerning its topology and shape, which constitutes a new application. Three methods are introduced to handle the optimization problem. The first one discusses the shape optimization of a simple predefined geometrical form. The second, studies the topological optimization based on bit array representation of a triangular mesh using genetic algorithms. An attempt to overcome the limitations of bit-array was developed through several steps (new partition element, mesh distinguishing between the representation type and finite element analysis, varying finite element problem size, creating a density vector to control the presence or absence of the boundaries, generating initial population representing void domains instead of void elements). The third method concerns shape optimization, based on a variable number of points creating a structural domain. In contrary to traditional methods where the variable points are indicated by the user, it searches for the optimum number of points to create the optimum shape. Finally, a comparison between these methods for the case of the floating breakwater is discussed.