Abstract: Objective To construct a model using radiomics to predict the mutation status of isocitrate dehydrogenase （IDH） in gliomas. Methods A retrospective analysis was conducted on preoperative magnetic resonance imaging （MRI） images of 336 glioma patients, and imaging features were extracted to construct an radiomics model. Receiver operating characteristic curve was used to evaluate the effectiveness of the model in predicting IDH mutation status. Results The radiomics features of each sequence were compressed and selected, of which 26 radiomics features were obtained from T₁WI, 24 radiomics features were obtained from T₂WI, 12 radiomics features were obtained from enhanced T₁WI, and 27 radiomics features were obtained from the three sequences combined. In the single sequence model, the area under the curve （AUC） of T₁WI was 0.780 （95%CI 0.724-0.836） in the training set and 0.763 （95%CI 0.650-0.876） in the test set. The AUC of T₂WI was 0.790 （95%CI 0.736-0.845） in training group and 0.785 （95%CI 0.677-0.893） in test group. The AUC of enhanced T₁WI was 0.815 （95%CI 0.762-0.867） in the training group and 0.810 （95%CI 0.702-0.918） in the test group. The radiomics model based on three sequences （T₁WI, T₂WI and enhanced T₁WI） combined had the best prediction performance. The area under the curve of the training group was 0.877 （95%CI 0.837-0.917）, and the area under the curve of the test group was 0.862 （95%CI 0.773-0.952）. Conclusion The radiomics feature model based on conventional MRI can effectively predict the IDH genotype of gliomas, thereby further guiding clinical diagnosis and treatment, and evaluating patient prognosis.