In the judicial practice of public security, seal paste is playing an important role in identifying the authenticity of documents for a long time. To provide evidence for determining the authenticity of the documents, a rapid nondestructive test was performed on 36 samples of inkpad using a SERDS Portable-Base portable differential Raman spectrometer under the test conditions of excitation source 785 nm, laser power 380 mW, scan time 10 s, and scan range 250~2 800 cm-1. The samples can be distinguished according to the position and number of characteristic peaks by analyzing the spectrum, and the samples in the same group can also be distinguished by the relative peak height ratio of characteristic peaks. At the same time, the principal component analysis method is used to reduce the dimension of the difference data, and the reduced dimension component matrix is systematically clustered. All the samples are scientifically divided into 7 categories, and the DP value is 85.23%. At the same time, a classification model based on cluster analysis is established by discriminant analysis. After cross validation, the classification accuracy is 94.4%. This method is fast, nondestructive, sensitive and strong anti fluorescence interference. Combined with the established model, it can scientifically test the inkpad samples, which can provide help for the front-line document inspection of public security.