The noise level in the cabin of a car has become one of the primary concerns for automobiles and cooling fans are one of the main noise sources. This paper investigates the source of broadband noise from a shrouded automotive cooling fan using a hybrid approach. Based on the surface pressure data obtained by CFD, the radiated noises are predicted using acoustic analogy. The accuracy of noise prediction and the reliability of the model analysis are verified by comparing both numerical and experimental results. The pressure information on the blade surface is analyzed to clarify the sources of broadband noise induced by the geometry-wake interaction. In future, these results can anticipate reducing the broadband noise in automotive cooling fans. (C) 2016 Elsevier Ltd. All rights reserved.