A commercially available computational fluid dynamics (CFD) code was tested for feasibility of application to automotive engine-cooling fans. Presence of the fan housing (shroud) and engine block makes the problem complex, as the computation has to deal with both stationary and moving frames of reference. One foreseeable benefit of CFD is the reductionæif not eliminationæof the number of prototype fans to make and test, in order to determine the design. Numerical experimentation could be carried out, in place of actual testing, to arrive at the optimum design. A couple of existing fans were chosen to assess the capabilities of the CFD code, and performance predicted. The computa- tional model is a fan set up on the fan wind tunnel. While predictions of the static pressure curve saw generally good agreement with measurements, those of fan torque suggested room for further improvement. Predicted velocity field around the fan was utilized to improve on the fan blade design; samples were made accordingly, and tested. Results confirmed the usefulness of the code.