The application of Support Vector Machines (SVMs) in face recognition is investigated in this paper. SVM is a classification algorithm recently developed by V. Vapnik and his team. We illustrate the potential of SVMs on the Cambridge ORL face database, which consists of 400 images of 40 individuals, containing quite a high degree of variability in expression, pose, and facial details. Our face recognition systems consist of two major phases. We present an automated facial feature extraction procedure and make use of Near set approach to choose the best feature among the considered one which significantly improves face recognition efficiency of SVM. Near Set approach was introduced by James Peters in 2006, as a result of generalization of rough set theory. One set X is near to another set Y to the extent that the description of at least one of the objects in set X matches the description of at least one of the objects in Y. Also we have shown that for face recognition in ORL face database using SVM with feature selection by near set approach has error rate 0.2% which is very less as compared to error rate obtained in the previous work done by other authors on the ORL face database.