The recognition of adulteration in fresh orange juice was performed by electronic tongue system. The raw data were analyzed by using principal component analysis (PCA),discriminant fact analysis (DFA) and partial least-squares analysis (PLS). DFA was more competent than PCA in distinguishing effect. It appeared regularity with the increasing proportion of adulterated samples. It had a high correlation between electronic tongue sensors response signals and the adulteration ratio of fresh orange juice, the correlation coefficient was 0.963 8, the prediction error of established PLS model ranged from 1.85% to 2.20%. It proved that electronic tongue could be applied in adulteration recognition of fresh orange juice.