Many of the finishing operations in the manufacture of dies and moulds are dependent on ball-nose milling cutters to produce the required geometry and surface quality. These cutters (high cost, carbide coated) work best in situations where the load on the cutter is constant and rapid changes in direction are kept to a minimum. The majority of computer aided manufacturing (CAM) systems provide a number of alternative machining strategies to support this approach, but they do not provide an accurate prediction of the surface quality in the finished product. This paper outlines a predictive theory for the surface quality. The ball-nose cutting mechanism has been modelled and simulated to generate the surface topography expected in the actual operation. The results match those from the experimental machining trials and enable the prediction and determination of the surface quality in other applications prior to the machining operations.