An artificial olfaction system based on visualization sensor array was employed to identify different ages of vinegar. In the image processing module of this system, the influence of different methods on the localization of the target image center was compared, including minimum enclosing rectangle, ellipse fitting and one-order moment. Since the target image was similar to the circle, all of the three methods could obtain center coordinates exactly, except that the last method consumed less time. Moreover, the characteristic region was reselected, which could better represent features of the target image. Usually, feature values are extracted based on the RGB color space. Then, each component and coordinate value in RGB, HSV and Lab color spaces were extracted and used as eigenvalues. The result showed that the data obtained from the Lab space had high stability. In order to identify different ages of vinegar, five different years of vinegar samples from 2011 to 2015 were selected in the experiment. The characteristic data from three kinds of color spaces was analyzed with principal component analysis (PCA) and linear discriminant analysis (LDA). Although the samples of vinegar in different years had a certain clustering tendency, especially in the Lab color space, there were still some samples overlapping each other and difficult to separate by PCA alone. Then this data was used as the input of LDA classifier for discriminate analysis. The recognition accuracy rate in the training set and testing set achieved 98% and 94% respectively in Lab color space, while the detection accuracies were not higher than 90% in other color spaces.