Researchers have been relentlessly trying to apply various techniques like KNearest Neighbor (KNN), Naïve Bayesian Classification and Artificial Neural Networks (ANN) for classifying the customer review documents that are abundantly available in the internet. In this article, an attempt has been made to use the Mahalanobis-Taguchi System (MTS) for the purpose of text classification. MTS was used to classify 300 opinions obtained from IMDB movie review data set with 150 negative opinions and 150 positive opinions. An amazing classification accuracy of 96.6% has been achieved using the proposed MTS based classifier. The classification performance of the MTS classifier was compared with other popular methods including Naïve Bayesian and Classification and Regression Tree (CART) analysis. © EuroJournals Publishing, Inc. 2011.