The gold content of published journals can reflect the academic level of a scholar, and the division of journals is an important index to reflect the quality of journals. The research on the prediction of journal division is helpful for scholars to quickly understand the level of journals, and can also provide relevant suggestions for the long-term development of journals. Using Clarivate Analytics’ 2022 Journal Citation Report, the author used the journal citation index, impact factor, immediate index, and subject ranking of journals as characteristic variables to predict the partitioning of SCI journals. One-vs-Rest algorithm is used to build multi-classification decision tree model, and grid search is used to optimize parameters. Finally, ROC curve and AUC value are used to evaluate the model. The results show that the AUC value of each category is above 0.85, and the feature importance ranking indicates that the citation index and impact factor have great influence on the journal partitioning. The algorithm and model in this paper perform well in the prediction of periodical partition, and can show the position of the periodical directly and the factors affecting the quality of the periodical. It provides a new way to broaden the prediction method of periodical classification.