Thirty-nine yellow cigarette case samples were tested by X-ray fluorescence spectrometry, Raman spectrometry, and differential Raman spectrometry to establish a classification method for cigarette case evidence. Using the hierarchical clustering method, the samples can be classified into five categories based on X-ray fluorescence spectrum data. The Raman spectrum data and differential Raman spectrum data are analyzed by synthesis. The principal component analysis is used to reduce the dimension of the data, followed by K-Means clustering to achieve more accurate classification based on hierarchical clustering. The samples can be classified into 12 categories. The results show that the proposed method fully combines the advantages of X-ray fluorescence spectrometry and differential Raman spectrometry, and combined with chemometrics, it can achieve better classification tests for cigarette case samples and provide clues and direction for the investigation work of public security organs.