Lycium barbarum(Lycii Fructus, wolfberry) is a valuable traditional Chinese medicinal material for medicine and food. Ningxia Zhongning wolfberry cultivars have been introduced to Gansu, Xinjiang, Qinghai and other provinces. Due to long-term introduction and improvement, there are many varieties of Lycium barbarum, one of the important factors affecting the quality of Lycii Fructus. This study used Fourier transforms infrared spectroscopy (FTIR) to determine the collected Lycii Fructus of Ningqi No. 1, No. 4, No. 5, No. 7, No. 0909 and No. 10 from Zhongning, Ningxia. The range of the one-dimensional infrared spectrum and the second derivative spectrum was 4 000~650 cm-1. After obtaining the spectrum, the infrared spectrums similarity coefficient was calculated using the fingerprint region (1 800~650 cm-1) with dense spectral bands. Then, the infrared fingerprints of different varieties of Lycii Fructus were classified and compared with SMICA cluster analysis. The results show that the one-dimensional infrared spectroscopy of different varieties of Lycium barbarum samples are relatively similar, with the location, peak height and peak shape of the peaks being relatively close. Their common absorption peaks are more numerous, and only the intensity, peak location and peak shape of the absorption peaks at around 3 282～3 288, 1 239～1 242 and 1 143～1 147 cm-1 are different, indicating that the polysaccharides, glycosides, proteins, lipids, and The types and contents of polysaccharides, glycosides, proteins, lipids, flavonoids and other components in different varieties of Lycium barbarum varied. In the second-order derivative spectra, the absorption peaks of NingQi 1, 4 and 7 were not apparentat 2 880 cm-1, and 0909 at 969 cm-1. The similarity of different varieties of Lycii Fructus ranges from 0. 9489 to 0. 9928, indicating certain differences among different varieties. Ningqi 7 has the lowest average similarity coefficient of 0. 9640, indicating that its component specificity is the highest. The similarity coefficient of No. 0909 and Ningqi 10 is 0. 992 8, the two varieties with the highest similarity. The cluster analysis was carried out using the Assure ID software with the absorption wave number of each medicinal material as a variable. The class spacing between Ningqi No. 1 and other varieties was small, ranging from 2. 17 to 2. 97. The class spacing between No. 0909 and other varieties was the largest, ranging from 2. 97 to 8. 06. In the clustering model, the recognition rate of different varieties of Lycii Fructus is 100%, only the rejection rate of Ningqi No. 1 is 66%, which can easy to be confused with other varieties of Lycii Fructus, and the rejection rate of No. 0909 is 100%, which is the easiest to distinguish. In the class model diagram, No. 0909 and Ningqi No. 5, No. 0909 and Ningqi No. 7 are separated in pairs, which can identify the samples of different varieties. The cluster analysis model was verified using known varieties of Lycii Fructus, and its recognition and rejection rates were the same as those of the cluster analysis model. The results of the one-dimensional infrared spectrum, second derivative spectrum, similarity, class spacing, recognition rate and rejection rate of different varieties of Lycii Fructus showed that Ningqi No. 1 was the original variety of other varieties of Lycii Fructus, and No. 0909 was significantly different from other varieties. Therefore, the combination of infrared spectrum and cluster analysis can quickly and nondestructively identify different varieties of Lycii Fructus, which has a certain guiding role in the breeding and production of Lycii Fructus in Ningxia.