The prosthesis design is a delicate and accurate engineering task that can be automated in the early steps before manufacturing. The challenge of the creation of the biomedical model is the complexity of the geometrical modelling as we have to deal with natural shapes. The representation of the human bones in terms of machining parameters the bottleneck in the design of complex products, and concurrent procedures can aid in this task. This work addresses the design requirements in order to build an anatomical skull prosthesis piece in the Computed Aided Design (CAD) systems. A novel methodology based on ellipse adjustment has been investigated in order to define the manufacture parameters. In geometric terms an ellipse seems with the bone's border shape in a Computed Tomography (CT) slice. The arc that fills the correspondent failure in the bone border is extracted from the respective adjusted ellipse to each CT slice and the set of those extracted arcs can be superimposed to define the stack of images to build a 3D CAD model. Evolutionary Algorithms were also applied to improve the quality of data generated. A prototype was implemented by an open source Java based tool (ImageJ) in order to create synthetic defects to simulate problems in the 3D virtual skull model. In context of product development this approach brings an essential integration between design and manufacturing process to reduce the elapse time among the medical procedure, modelling and machining.