Abstract: A fuzzy neural network for identifying 11 kinds of mineral waters is developed based on an entropy based clustering method Partitioning fuzzy input space and extracting fuzzy IF THEN rules are implemented employing the clustering method and the Gradient Descent algorithm is used for optimizing system parameters, so that the system has good interpretability and learning capability Experimental results show that the system is feasible and effective for identifying 11 kinds of mineral waters by its taste signals