Objective：To investigate the value of radiomics based on dynamic contrast-enhanced magnetic resonance imaging（DCE-MRI） in predicting luminal and non-luminal breast cancer. Methods：A retrospective analysis was performed for the clinical data of 42 patients who underwent breast DCE-MRI and had pathologically confirmed breast cancer. The lesions were outlined by manual segmentation on the images with the strongest enhancement of DCE-MRI to determine radiomic features. Then the least absolute shrinkage and selection operator was used to select the most important features. All patients were divided into luminal group and non-luminal group according to the expression of estrogen receptor（ER） and progesterone receptor（PR）. Classifiers were constructed based on logistic regression（LR），random forest（RF），K nearest neighbor（KNN），and support vector machine（SVM），and the 5-fold cross validation method was used to verify prediction performance. Results：A total of 4 optimal features were screened out，consisting of 1 shape feature and 3 wavelet features，among which Original_Shape_Maximum 2D Diameter Row，Wavelet-LLH_GLCM_Idn，and Wavelet-HHH_GLCM_Correlation had significant correlations with ER and PR levels. The RF-based classifier had the best performance in predicting luminal and non-luminal breast cancer，with a sensitivity of 0.838，a specificity of 0.900，an accuracy of 0.853，and an area under the ROC curve of 0.876. Conclusion：Radiomics based on DCE-MRI has a good value in predicting luminal and non-luminal breast cancer，and the RF-based classifier has the best performance.