Abstract: A characteristic spectrum of Gastrodia elata was established based on the crossing response of gustatory system of mammals to identify true and false Gastrodia elata. This detection system was simplified. The array sensor with 8 sensing units was established by a piece of 96-well plate.Signal were collected by the ELIASA and combined with signals of 8 sensing units into the characteristic spectrum of detection objects. Gastrodia elata, fresh Gastrodia elata and four common categories of fake Gastrodia elata ( Canna edulis Ker, potato, Dobinea delavayi and Radix Mirabilis) were identified by this sensor.Data processing was carried out by combining principal component analysis ( PCA) , hierarchical clustering analysis ( HCA) and linear discriminant analysis ( LDA) . PCA results demonstrated that this sensor mainly identifies 6 samples based on polarity and p H value.According to HCA results, 24 samples of 6 categories were all classified accurately. LDA results showed that the identification accuracy of this sensor to six categories reached 100%. The established array sensor could be simplified into a sensor with 3 sensing units by analyzing correlation of 8 sensing units. The simplified sensor also can make accurate classification of 6 sample categories.The proposed method is simple and effective and can make mass-operations. It possesses promising potential application values in fast flux identification of Gastrodia elata.