In order to solve the problem that the current multi-label classification algorithm has low classification accuracy and high computational complexity, a centroid-based multi-label model for text categorization, named Multi-label Gravitation Model (ML-GM), is proposed. In the training phase, a similarity interval by calculating the similarity between the document and the centroid of the class. In the test phase, multi-label classification is performed by comparing the similarity between the undefined document and the class centroid is within the similarity interval. The model solves the problem of high computational complexity and low classification accuracy by introducing a centroid classifier and a gravity model. The Yahoo dataset is used in the experiment, and the results show that ML-GM achieves supe- rior performance in terms of average accuracy, AUC, one-error and hamming loss.