This retrospective analysis focused on the potential of CT-based radiomics to improve the noninvasive prediction of HER2 status in gastric cancer, which is essential for guiding personalized treatment strategies. Clinical variables, including pathological T stage, N stage, and carcinoembryonic antigen levels, were identified as significant predictors through multivariate analysis. In parallel, several radiomics models were developed from CT images, with a support vector machine demonstrating the best performance. A combined model integrating clinical features with the radiomics score achieved superior predictive accuracy compared to either model alone, showing robust performance in both the training and testing cohorts. Overall, these findings highlight the added value of combining CT-based radiomics with clinical data to achieve a more accurate and noninvasive approach to HER2 status prediction in gastric cancer.