Neurosurgery interventions involve complex tracking systems because a tissue deformation takesplace. The neuronavigation system relies only on preoperative images. In order to overcome the soft tissue deformations and guarantee the accuracy of the navigation a biomechanical model can be used during surgery to simulate the deformation of the brain. Therefore, a mesh generation for an optimal real-time Finite Element Model (FEM) becomes crucial. In this work we present different alternatives from a meshgeneration point of view that were evaluated to optimize the process in terms of elements quantity and quality as well as constraints of a intraoperative application and patient specific data.