Abstract: Two kinds of Xinyang Maojian tea ( Mopanshan and Shi Hekou, each had five grades) were detected by an electronic tongue.Multivariate statistical data analysis techniques as principal component analysis ( PCA) and partial least squares ( PLS) discriminant analysis were used to classify and identify the grade of samples. Four grades of Mopanshan tea and Shi Hekou tea were clearly classified on the PCA plot, and the discrimination indexes were 78.94 and 84.22, respectively.Two PLS models with good correlation indexes ( > 0.9) were built, and the predicting results of unknown samples by the models coincided with their actual grades.The results obtained in this study indicated that the electronic tongue could be efficiently used to predict the quality grade of tea.