Efficient method for differentiation and identification of pathogenic bacteria was important to ensure food safety and human health. Infrared spectral technology has the advantages of comprehensive information， high resolution， and fast identification. In this study， FT-IR fingerprint absorption spectra of three subspecies of Salmonella paratyphi were collected， and a standard spectral derivatives library was created. Combined with chemometrics methods， two cluster models of PCA and HCA were established. It was found that three subspecies of Salmonella paratyphi could be well differentiated and identified. Spiked assays proved that the standard spectral derivatives library and the cluster models could be used to differentiate and identify the three subspecies of Salmonella paratyphi. As a rapid， easy-to-use， accurate， and low cost technique， FT-IR spectroscopy is an effective tool to differentiate and identify three subspecies of Salmonella paratyphi.