Objective: Efforts should be made to achieve complete cytoreduction during interval debulking surgery(IDS) after neoadjuvant chemotherapy(NAC) to treat ovarian cancer, due to optimal cytoreduction being an important prognostic factor. This study aimed to create a new model to predict optimal cytoreduction using computed tomography (CT) imaging before IDS. Methods: Preoperative CT scans of 72 patients who underwent IDS were retrospectively reviewed. We classified the abdominal cavity according to Sugarbaker's peritoneal cancer index (PCI) and scored the tumor size and degree of tumor invasiveness from 0 to 5 points for each CT feature. Residual disease measuring < 0.5 cm in maximal diameter after IDS was considered optimal cytoreduction. We developed a model using each scores to predict the residual disease, logistic regression and random forest were employed. A new model was created using the random forest that has been confirmed to have higher predictive power. The performance of the novel model has been reported, and the correlation between residual disease and progression-free survival (PFS) has been assessed. Results: The rate of optimal cytoreduction in the total study population was 59.7% (n = 43). Patients with optimal cytoreduction after IDS had significantly longer PFS than other patients (p = 0.04). CA125 levels before IDS did not affect residual disease (Area under the ROC curve (AUC) = 0.584, 95% CI: 0.450-0.719). Multivariable analysis resulted in a prediction model that included disease features of greater omentum, ascending colon and right paracolic gutter invasion with an AUC of 0.651 (95% CI: 0.539, 0.763). Using random forest, the top three CT features were selected with a threshold of 0.1; the greater omentum, pelvis, and lesser sac/lesser omentum. The top three features achieved an AUC of 0.729 (95% CI: 0.622-0.833). Conclusion: A low CT disease score of disease at top three CT features can be a strong predictor for optimal cytoreduction in IDS using out model.