Abstract: In order to realize the rapid identification of sesame oil varieties,this paper proposed a method for sesame oil variety identification based on terahertz time domain spectroscopy(THz-TDS). The spectrum in the range of 0~2.5 THz was selected for analysis,and the time domain spectral data was reduced by principal component analysis(PCA). The first four principal components(cumulative contribution rate was over than 99%)were selected to represent the original data,and then the support vector machine(SVM)method was used to classify and identify different varieties of sesame oil. Three different kernel functions were used for classification,and the grid model was used to obtain the optimal model and its model parameters. Using the radial basis kernel function(parameter penalty function C=0.01,kernel function coefficient γ=0.1),the model recognition rate was the highest,reaching 100%. It indicated that the terahertz time domain spectroscopy technology combined with PCA and SVM methods could be quickly and reliably consumed,providing a new technical means for the identification of food safety.