This paper addresses a novel methodology of extracting the human emotions by considering Skew Gaussian mixture model. The features of the emotions will affect the recognition efficiency of the speech recognition systems. Various techniques are used in identifying the emotions. The proposed system is experimented over a gender independent emotion data base. In order to extract the emotion features like pitch, energy and frequency range features are considered. The proposed system is tested using a Synthetic speech data set together with the standard data set of Berlin. This model is evaluated in the presence of noise and without noise the efficiency of the model is evaluated and presented by using confusion matrix. The results are tabulated in presented using a confusion matrix. © Research India Publications.