Static Security Assessment (SSA) is a major concern in planning and operation stages of electric power systems. The traditional method used in static security analysis involves solving full AC load flow equations for each contingency. This is highly time consuming and inadequate for real time applications. The Pattern Recognition (PR) approach is recognized as an alternative tool for on-line security evaluation. This paper proposes a recently introduced machine learning tool called Support Vector Machine (SVM) in the classification phase of pattern recognition approach. Many feature selection algorithms are used for selecting optimal feature subset in the design of PR system. The proposed SVM based PR approach is tested on IEEE 14 Bus and IEEE 57 Bus Systems. The performance of SVM classifier are compared with other classifiers like Multilayer Perceptron, Method of Least Squares and Linear Discriminant Analysis classifiers. Simulation results prove that SVM classifier gives a fairly high classification accuracy and less misclassification rate compared to other equivalent classifier algorithms. Copyright © 2010-11.