The chemical composition and ratio of ancient glasses are highly susceptible to change due to the complex burial environment. In order to determine the classification rules of lead-barium glass and potassium glass, and to subclassify lead-barium glass and potassium glass according to the appropriate chemical composition index, this paper mines the linear combination of high potassium glass and lead-barium glass by linear discriminant analysis, and calculates the clustering centers of lead-barium glass and high potassium glass subclassifications and the subclassifications of glass to be classified based on the FCM (Fuzzy-C Mean Clustering) algorithm of genetic simulated annealing algorithm. The sub-clusters of lead-barium glass and high-potassium glass, as well as the affiliation degrees of the glass samples to be classified were calculated based on the genetic simulated annealing algorithm. The model results were then examined by using the decision tree algorithm, and the results of glass artifact type identification were obtained.