In recent years, the increase in the computing capacity of modern machines has led to the production of ever more realistic digital graphic elements. At the same time, the possibility of automating increasingly complex processes, has led the research to have faster, safer and more reliable results in every field. On medical field, one of the most significant application of the innovations described above, concerns the reconstruction of 3D facial models from 2D DICOM data. This paper analyzes the development of an automated solution for the reconstruction of 3D models of soft tissue and hard bones, to support maxillofacial surgery. The solution presented, starting from 2D medical data in dicom format, comes to the creation of a 3D model, applying segmentation techniques to divide the soft tissue from the hard bones, and reconstructying the 3D model using an algorithm of marching cubes.