This paper considers the influence of the turbulence model adopted to numerically predict the aero-elastic phenomena in Fluid Structure Interaction (FSI) problems. After a detailed evaluation of the main suitable turbulence models available in the Reynolds Average Navier Stokes (RANS) approach, the two-equation k-ω turbulent model was identified as the best compromise between solution accuracy and computational cost. In order to test the k-ω turbulence model performance for FSI applications, a specific simulation methodology was defined. It was firstly tested in assessing the aerodynamic forces resulting on static bluff bodies for which experimental databases were available. Then, the simulation methodology based on the k-co turbulence model was used to reproduce the steady and unsteady aerodynamic forces induced on a Great Belt East bridge cross section subjected to wind load. In particular, the unsteady aerodynamic forces induced on the bridge were evaluated by calculating the flutter derivatives. As for the test cases, also for the Great Belt East bridge the computational results were compared to experimental evidence obtained by wind tunnel tests.