Development and validation of a radiomics nomogram for predicting recurrence and adjuvant therapy benefit populations in high/intermediate-risk gastrointestinal stromal tumors (GISTs) based on computed tomography (CT) radiomic features. Retrospectively collected from 2009.07 to 2015.09, 220 patients with pathological diagnosis of intermediate- and high-risk stratified gastrointestinal stromal tumors and received imatinib treatment were randomly divided into (6:4) training cohort and validation cohort. The 2D-tumor region of interest (ROI) was delineated from the portal-phase images on contrast-enhanced (CE) CT, and radiological features were extracted. The most valuable radiological features were obtained using a Lasso-Cox regression model. Integrated construction was conducted of nomograms of radiomics characteristics to predict recurrence-free survival (RFS) in patients receiving adjuvant therapy. Eight radiomic signatures were finally selected. The area under the curve (AUC) of the radiomics signature model for predicting 3-, 5-, and 7-year RFS in the training and validation cohorts (training cohort AUC = 0.80, 0.84, 0.76; validation cohort AUC = 0.78, 0.80, 0.76). The constructed radiomics nomogram was more accurate than the clinicopathological nomogram for predicting RFS in GIST (C-index: 0.864 95%CI, 0.817-0.911 vs. 0.733 95%CI, 0.675-0.791). Kaplan-Meier survival curve analysis showed a greater benefit from adjuvant therapy in patients with high radiomics scores (training cohort: p < 0.0001; validation cohort: p = 0.017), while there was no significant difference in the low-score group (p > 0.05). In this study, a nomogram constructed based on preoperative CT radiomics features could be used for RFS prediction in high/intermediate-risk GISTs and assist the clinical decision-making for GIST patients.