Neural Networks (NN) have been applied to the security assessment of power systems and have shown great potential for predicting the security of large power systems. The curse of dimensionality states that the required size of the training set for accurate NN increases exponentially with the size of input dimension. Thus, an effective feature reduction technique is needed to reduce the dimensionality of the operating space and create a high correlation of input data with the decision space. This paper presents a new feature reduction technique for NN-based power system security assessment. The proposed feature reduction technique reduces the computational burden and the NN is rapidly trained to predict the security of power systems. The proposed feature reduction technique was implemented and tested on IEEE 50-generator, 145-bus system. Numerical results are presented to demonstrate the performance of the proposed feature reduction technique.