Seventy-six samples of ink pads and ink oils of different brands and types in the domestic market were selected. Under simulated actual conditions, they were evenly stamped on the same type of A4 printing paper with normal force to obtain 76 samples of ink color materials. The fluorescence images of the samples were collected using the Dahao-VF10 Pro ultra-deep field high-magnification video fluorescence microscope. The average pixel values of RGB and LCH of the fluorescence images of the samples were calculated through Matlab. The T-SNE algorithm was used to perform K-means clustering analysis on the brightness and color values of the fluorescence images of the samples, and then comprehensive quantitative inspection and identification were carried out. The research results show that 76 kinds of ink pads and ink oils can be classified into 4 categories based on the difference of LCH values for brightness clustering analysis, and into 6 categories based on the difference of RGB values for color clustering analysis. Through comprehensive analysis of fluorescence brightness and color, 76 kinds of ink pads and ink oils can be ultimately classified into 21 categories. Video fluorescence spectral image technology can widely and effectively conduct systematic analysis and effective identification of ink pads and ink oils of different brands and types, providing a new technical means with high practical value for the forensic science inspection of seal impressions.