In the present work, the authors explored a rapid method of the Zanthoxylum nitidum geographical origins determination. Based on Fourier transform infrared spectroscopy (FTIR) technology, the band of 1 800～400 cm-1 which is the IR fingerprint of Zanthoxylum nitidum, the Fisher ratio and the soft independent modeling of class analogies (SIMCA) were used to build a classification model. Respectively, four kinds of Zanthoxylum nitidum in the Guangxi region were detected by the model, and the model was verified by calculating their recognition rate and rejection rate. The results show that the authors can accurately extract the overall information of Chinese herbal medicines by using the FTIR, also established a pattern recognition model to predict unknown samples, and obtained satisfactory recognition rate and rejection rate, indicating that the model has stronger ability of identification. The detection on real time was carried out rapidly with the Fisher model, suggesting that the model has more practical value.