Fourier transform infrared spectrometer was used to collect infrared spectra of Cortex Phellodendri from six different regions. Original spectra were preprocessed by carrying out appropriate baseline correction and five-points smoothing, and the averaged spectra of Cortex Phellodendri from the six origins were analyzed. As a result, the averaged spectra looked quite similar. The normalized spectra were selected to construct principal component analysis model in the range of fingerprint region 1 800～500 cm-1, and according to the model, the first three principal components accounted for 98% of the variance information in the fingerprint region, and each sample was able to form distinct cluster in the principal component space, then the identification of Cortex Phellodendri from the six regions was basically achieved; besides, to some extent, the sparse density of the samples distribution reflected the genetic relationship. The loading factors of the model were analyzed, and the results indicated that the differences between Cortex Phellodendri samples mostly depended on the contents of protein, carbohydrates, lipids, alkaloids, sterols, obaculactone, oba-cunone, and obacunonlc acid. On the whole, combined with principal component analysis, FTIR provides an effective way to evaluate the herbal Cortex Phellodendri rapidly and nondestructively, which also reflects the content difference of material composition.