Terahertz time domain spectroscopy (THz-TDS) combined with principal component analysis-linear discriminant analysis (PCA-LDA) and support vector machine (SVM) was used for identification of official rhubarb samples. Terahertz time domain transmittance spectra of 41 official and unofficial rhubarb samples were measured in time domain and then were transformed to absorption coefficients in frequency domain. Qualitative classification models of PCA-LDA and SVM were established based on the absorption coefficients and cross validated for identifying official and unofficial rhubarb samples. The predictive ability and stability of the models were evaluated using bootstrapped Latin-partitions method with 50 bootstraps and 4 Latin-partitions. Satisfactory results were obtained by using both PCA-LDA and SVM. The proposed method proved to be a convenient, non-polluting, accurate, and non-chemical treatment approach for identifying rhubarb samples. The developed procedure can be easily implemented for quality control in other herbal medicine classification and production.