Abstract In this work a novel electronic tongue (ET) using Fourier transform impedance spectroscopy has been demonstrated. Odd random phase multisine waveform has been used as an excitation signal. Texas Instruments’ PCM2900B USB audio CODEC chip has been used as a signal generation and an acquisition module for the ET. The acquired impedance features have been further subjected to Principal Component Analysis (PCA) for dimensionality reduction and Support Vector Machines (SVM) for pattern classification. A good classification accuracy has been achieved for single specie samples (taste samples), multi-species samples (water) and complex samples (tea). Also, the performance of the proposed system has been measured in terms of qualitative performance parameters namely, false positive rate, false negative rate, sensitivity rate and specificity rate.