This paper presents a model for composition analysis and identification of glass products. Based on the XGBOOST algorithm, the XGBOOST model was constructed to find the type of glass, ornamentation, and color in the weathering and non-weathering points. The main component of glass products is SiO2, which imparts the desired functional properties by mixing different oxides. Ancient glass products are very susceptible to weathering caused by the buried environment, resulting in changes in the proportions of the various components inside. Since glass products of different materials have different properties, the classification of glass products is worth studying.