Facial Expressions of the human being is the one which is the outcome of the inner feelings of the mind. It is the person�s internal emotional states and intentions.A person�s face provides a lot of information such as age, gender, identity, mood, expressions and so on. Faces play an important role in the recognition of the expressions of persons. In this research, an attempt is made to design a model to classify human facial expressions according to the features extracted f0rom human facial images by applying 3 Sigma limits inSecond level decision using Neural Network (NN). Now a days, Artificial Neural Network (ANN) has been widely used as a tool for solving many decision modeling problems. In this paper a feed forward propagation Neural networks are constructed for expression classification system for gray-scale facial images. Three groups of expressions including Happy, Sad and Anger are used in the classification system. In this paper, a Second level decision has been proposed in which the output obtained from the Neural Network(Primary Level) has been refined at the Second level in order to improvise the accuracy of the recognition rate. The accuracy of the system is analyzed by the variation on the range of the expression groups. The efficiency of the system is demonstrated through the experimental results.