Abstract: In this research, an electronic tongue was employed to distinguish goldfish’s cultured water. Three pattern recognition methods were used, which were principal component analysis (PCA), cluster analysis (CA) and canonical discriminant analysis (CDA), and different cultured density and different cultured day were distinguished by data obtained by e-tongue and physico-chemical index. By results comparison of these three methods, different cultured density and different cultured day could be better distinguished by CDA, and could also be distinguished by PCA combined with physico-chemical index. However, different cultured water could not be distinguished by CA effectively. The results also showed that e-tongue can be employed to distinguish different cultured water by CDA and PCA.