In order to construct a model for identifying printing and copying toners, X-ray fluorescence spectroscopy was used for theoretical analysis and experimental verification. The experiment data was obtained through quantitatively analyzing of the element content of 28 common brands toner samples. According to the preliminary grouping based on the presence or absence of specific elements, improved power k-means algorithm was used to perform clustering analysis to divide samples into 7 groups. After the verification by the Pearson coefficient, the result was testified, and the samples were well distinguished. The samples within the group were significantly correlated at the level of 0.001.The experiment provides new ideas which is fast and easy for the identification of printing and copying toners in judicial practice.