To explore the performance of MR radiomics in predicting the pathological classification and staging of thymic epithelial tumors (TETs), we built two radiomics models based on support vector machine. Besides, we developed a radiomics nomogram for predicting risk stratification of advanced TETs. The models achieved an area under the curve of 77.1% or 90.8% in the test cohort in distinguishing low-, high-risk thymomas and thymic carcinomas or early and advanced TETs. The radiomics model, symptom, and pericardial effusion constituted a radiomics nomogram, with a C-index of 0.957 in the test cohort. Thus, MR radiomics can be useful for assessing TETs.