This study aims to explore the use of spectrophotometry and hyperspectral imaging technology combined with machine learning to effectively distinguish the types of stamp pad ink. Hyperspectral data and chromaticity values of 27 stamp pad ink samples from different brands and models were collected. Principal component analysis (PCA) was applied to the average chromaticity data for dimensionality reduction, and K-Means cluster analysis was used to successfully classify the ink samples into four categories. Subsequently, four classification models, namely LightGBM (light gradient boosting machine), XGBoost (extreme gradient boosting), SVM (support vector machine) and KNN (K-nearest neighbor) were used. The test set and training set were determined at a ratio of 1:4, and the samples of each category in the results of cluster analysis were identified one by one. The results showed that the SVM model, LightGBM model, and XGBoost model performed well. Specifically, SVM achieved 100% accuracy for sample classification in categories Ⅰ, Ⅱ, and Ⅲ, and 98.3% for sample Ⅳ. This study provides a new method for quickly and accurately identifying the types of stamp pad ink.