In neurosurgery, effects of deformation should be considered to avoid damaging brain tissues. The goal of this study is to develop an automatic path planner considering the deformation of brain tissues. This paper shows a scheme which combines FEM (Finite Element Method) and an optimization method for optimization of retraction in order to approach a deep part of a brain. Also, evaluations of two optimization results are discussed. One optimization is for retraction of a simple shape model for comparing two solvers, Pattern Search and Genetic Algorithm. Pattern Search Algorithm obtained maximum view size for the simple model when the principal stress of the tissue is not more than the threshold 500 (Pa). The other optimization is for retraction of a brain fissure model. Based on the result of the simple shape model, Pattern Search Algorithm is used for this optimization. It successfully generated optimal position and posture of a spatula for opening the fissure model which has same mechanical property with the human brain. These results show the effectiveness of the proposed scheme.