Fan noise can form a significant part of the vehicle noise signature and needs hence to be optimized in view of exterior noise and operator exposure. Putting together unsteady CFD simulation with acoustic FEM modeling, tonal and broadband fan noise can be accurately predicted, accounting for the sound propagation through engine compartment and vehicle frame structure. This paper focuses on method development and validation in view of the practical vehicle design process. In a step by-step approach, the model has been validated against a dedicated test-set-up, so that good accuracy of operational fan noise prediction could be achieved. Main focus was on the acoustic transfer through the engine compartment. The equivalent acoustic transfer through radiators/heat exchangers is modeled based on separate detailed acoustic models. The updating process revealed the sensitivity of various components in the engine compartment. Unsteady CFD included the build-up of a sliding mesh model which was analyzed using the DDES method. After convergence, time data of blade surface pressure were exported in CGNS format. These pressure data were used to generate rotating dipole sources in acoustic FE analysis and predict the fan noise response in frequency domain at two selected rpm. Post processing includes the predicted noise at target microphone positions as well as colormaps of sound pressure distributions that can guide to the development of countermeasures.