The oil stained vinegar capsule and the qualified vinegar capsule are of the same color, which makes it difficult to detect oil stained vinegar capsule using the naked eye or computer vision. The spectral information of hyperspectral data is sensitive to the chemical compounds of sample area, and allows quantitative/qualitative analysis of biological products. Therefore the feasibility of using hyper-spectral imaging technology for rapid determination of oil stained vinegar capsule was investigated. The hyperspectral image data of oil stained vinegar capsule were acquired in the wavelength range of 430~960nm, and were used to extract characteristic information of oil stained vinegar capsule. Linear discriminant analysis (LDA) and K-nearest neighbor algorithm (KNN) were used to build discriminant models for oil stained vinegar capsule. When principal component factors equaled to 2 and K levels equaled to 3, the optimal KNN model was obtained with identification rates of 100% in both training set and prediction set. The overall results show that hyper-spectral imaging technology could extract the spectral characteristics of oil stained vinegar capsule efficiently. The hyper-spectral imaging technology could be used for rapid determination of oil stained vinegar capsule.