The study developed a computed tomography-based radiomic model to assess the incidence of extrapulmonary organ involvement and predict recovery time in children with Mycoplasma pneumoniae (MPP) pneumonia. They retrospectively analyzed 556 patients from three centers between October 2022 and December 2024. They used LASSO normalization and logistic regression to select parameters for disability assessment, with performance assessed via AUC, calibration and decision curves. The integrated model achieved the highest accuracy (AUC = 0.94; 95% CI, 0.84–0.99), which was statistically significantly better than the radiomic model, the clinical-laboratory model (AUC = 0.67; 95% CI, 0.49–0.84), or the imaging features model (AUC = 0.73; 95% CI, 0.53–0.88). For predicting recovery duration, the model was evaluated with a modified MSE of 6.0, MAE of 1.9, and R² of 0.6825. Radiomics has improved the accuracy of clinical and imaging parameters, providing clinicians with a tool to better diagnose MPP.