This paper proposes a reasonable methodology applied in classification and quantification techniques based on the voltammetric electronic tongue. We designed voltammetric electronic tongue oil sample pretreatment system with petroleum ether organic solvent. Through three-electrode system and cyclic voltammetry method processing blending soybean oil sample to get waveform output. Extracting crest methods as its feature extraction Pattern recognition use kernel principal component analysis and factor analysis blending on a different level of soybean oil, Partial Least Square (PLS) techniques was applied for data management and prediction models building, the prediction models are the blending ratio, the results show that voltammetric electron tongue can distinguish the quality of soybean oil.