The conventional method to identify tree peony varieties is based on morphology, which is strongly subjective and lacks chemical information. Fourier transform infrared spectroscopy (FTIR) technology combined with principal component analysis (PCA) is applied to identify different varieties of tree peony flowers in this paper. The infrared spectra of tree peony samples are similar on the whole, which are mainly composed of the vibration absorption bands of proteins, phenols, lipids, glycosides, polysaccharides, flavonoids, and so on. First derivative sepctra in the range of 1800~700 cm-1 are selected to perform principal component analysis and correlation analysis using Matlab 2010 program. Correct classification amounting for 96% is observed. The results of principal component analysis and correlation analysis are identical with each other. It is demonstrated that FTIR-principal component analysis can be used as a possible means for the identification of different tree peony varieties.