Abstract: Objective To develop and validate a multidimensional data-based prediction model for evaluating the prognosis of children with bronchial asthma following inhaled corticosteroid (ICS) therapy.Methods A total of 213 children with bronchial asthma, who were admitled to Suqian First People′s hospital from January 2021 to May 2025 and received standardized ICS treatment were enrolled and their clinical data were retrospectively analyzed. The Boruta algorithm was applied to identify key predictive variables from indicators related to inflammation, immunity, pulmonary function, and allergy. A logistic regression model was established, and the dataset was randomly divided into a training set and an internal validation set at a ratio of 7:3. Model performance was assessed in terms of discrimination, calibration, and reclassification ability using the net reclassification improvement (NRI) and integrated discrimination improvement (IDI) indices.Results Seven key predictive factors were identified by the Boruta algorithm, including interleukin-6 (IL-6), CD4⁺/CD8⁺ ratio, high-sensitivity C-reactive protein (hs-CRP), percentage of predicted forced expiratory volume in one second (FEV1%pred), vitamin D level, fractional exhaled nitric oxide (FeNO), and positivity for three or more allergens. The multidimensional integrated model demonstrated the optimal prediction performance, with an area under the receiver operating characteristic curve (AUC) of 0.971 in the training set and 0.948 in the validation set. The model also showed good calibration and significantly outperformed the conventional immune-inflammatory model in overall performance. Among the selected predictors, FeNO, the CD4⁺/CD8⁺ ratio, and IL-6 were the main prediction factors.Conclusions The prognostic prediction model incorporating multidimensional data, including airway inflammation, immune status, pulmonary function, and allergic burden, can accurately predict the risk of adverse prognosis in children with bronchial asthma receiving ICS therapy. The nomogram derived from this model may serve as a practical tool for clinical risk stratification and individualized intervention.