The present report deals with the aeroelastic response of high-rise buildings. The wind flow around an assumed square cylindrical high-rise building is modeled using the commercial CFD-program Ansys CFX 11. To model the Fluid-Structure Interaction (FSI) a FEM beam model is used to represent the structure while the dynamic model is set up using a modal representation. A method for generating 3D meshes suitable for exterior flow around a rectangular cylindrical building is presented, with basis in an algorithm for generating 2D meshes using a hyperbolic grid generation scheme. Some simple mechanisms to ensure a smoother mesh are presented as well. The difference between using structured and unstructured meshes are analyzed, and it is concluded that use of structured meshes is superior with respect to computation time. An analysis is then performed to see, which methods for governing mesh stiffness during mesh deformations in Ansys-CFX is best, and on basis of 2D analyses it is concluded that the use of the reciprocal of the wall distance as stiffness parameter, or use of the built-in function ’Increase Near Boundaries’ show superior performance. Due to limited availability of computational power in the project, analyses are made determining the effects of using simulation settings that are less than ideal. In these analyses it is shown that these delimitations show little influence on the loads on the building in the streamwise direction, but that the loads in the cross stream direction show strong dependency. The effect of modeling the aeroelastic response of a high-rise building is determined, by comparing the response of two different simulations, one where the structure is allowed to move freely, and one where it is stationary. It is concluded that there is an obvious difference between the two methods. Finally the aeroelastic load response of a high-rise building are obtained by use of flutter derivatives, and it is shown that this method gives qualitatively good results, and seems applicable to use on high-rise buildings. The response is found using the two first eigenmodes only, and the effects of including more eigenmodes is not analyzed.