〔Abstract〕 Objective To establish a radiomics method based on three-dimensional magnetic resonance imaging (3D-MRI) to predict the molecular typing of isocitrate dehydrogenase (IDH) in adult diffuse gliomas, so as to assist preoperative decision-making and prognosis assessment. Methods MRI images, clinicopathological and IDH molecular typing data of 66 adult patients with diffuse glioma confirmed by surgical resection and pathology in the First Affiliated Hospital of Shenzhen University from July 2016 to January 2021 were retrospectively collected. IDH molecular typing was performed by quantitative polymerase chain reaction (qPCR) sequencing. 3D regions of interest (ROI) were delineated based on T1WI、T2WI images of patients, and their imaging features were extracted. The data were randomly divided into training set and test set at 7:3. Spearman correlation analysis and LASSO were used to reduce the dimensionality of the features, and the radiomics features related to IDH molecular typing were screened out. The IDH predictive model was established and the performance of the prediction model was verified by ROC curve and confusion matrix. Results A total of 66 cases of adult diffuse glioma were included in this study, including 27 IDH mutant cases and 39 IDH wild type cases. 396 features were extracted from the images of each patient, and 6 related features were screened out to establish IDH prediction model. The area under curve (AUC) of the training set was 0.882. The AUC of the test set was 0.841. The accuracy, sensitivity and specificity of the training set were 0.894, 0.842 and 0.929, respectively. The accuracy, sensitivity and specificity of the test set were 0.789, 0.750 and 0.818, respectively. Conclusion MRI-based imaging can predict diffuse glioma in adult brain and obtain IDH classification results in a non-invasive way, which has certain clinical value for preoperative and inoperable patients to make diagnosis and treatment plans and follow-up strategies.