In this work, I present a method for stereo-based 3D face reconstruction. The algorithm is based on depth estimation of the face by using image pair which are taken from different point of views. Afterwards, surface mesh of the face is generated by deploying this depth information. In order to improve the depth estimation, I take advantage of the face properties.