After COVID-19 Aeronautics industries are at the cutting edge for reducing emission in order to stick with the ACARE 2030 objectives. Film cooled turbines are one of the main technology to achieve this objective and an accurate flow prediction is necessary. RANS simulations are nowadays the main tool used for performance prediction of turbines and aerodynamic components. Film cooled turbines geometries are very complex and as a result is almost impossible to mesh them with classical multi-block hexahedral meshing methods. For this reason flexible tetrahedral meshes are preferable. Discretization error (along with geometric error and modeling error) strongly impacts the evaluation of performances, leading to inaccurate results. Furthermore, the simulation process relies heavily on the experience of the engineer and it is not automatized. In order to automatize the simulation process, take advantage of the flexibility of the tetrahedral elements and in order to contain the discretization error, metric-based anisotropic mesh adaptation appears to be a suitable technology [1]. The applicability of the metric based anisotropic mesh adaptation to the TATEF2 film cooled turbine Nozzle Guide Vane (NGV) [3] has been shown in [2]. The goal of this paper is to analyse the benefit of mesh adaptation to the aerothermic performance prediction of the NGV. Different methods to compute the gradient at the wall will be tested in the framework of the Mixed Element Volume (MEV) RANS solver Wolf. The aerothermic performance coefficients will be compared with the experimental results and the impact of the remeshing chain will be evaluated.