The work aims to establish a method for non-destructive identification of white shopping paper bags by differential Raman spectroscopy. 60 white shopping paper bags of different brands and specifications were collected and detected by Raman spectroscopy. The Raman spectra of the samples were preprocessed, and the samples were preliminarily classified based on the spectra. The samples were grouped by stoichiometric methods. In addition, Fisher discriminant analysis method was applied to verify the classification results. Finally, the RBF model was applied to classify and distinguish unknown samples. Based on the difference in calcium carbonate, talc powder, and barium sulfate contained in the samples, the white shopping paper bag samples could be preliminarily divided into five categories. K-means clustering method was used for further subdivision, and Fisher discriminant method was used to verify the sample classification results, with a discriminant accuracy of 100%. The application of neural network RBF model for discriminative analysis of unknown samples achieved an accuracy of 89.48%. This method is simple and easy to implement, which provides a scientific basis for the classification of white shopping paper bags, and also provides a convenient way for the development of grassroots public security work.