Mycoplasma pneumonia (MP) is a relatively common infection in children. While sequential treatment with azithromycin is a commonly used regimen, therapeutic response varies substantially among children. This study aims to establish a column chart prediction model based on the clinical characteristics and pathogenic outcomes of Mycoplasma pneumonia in children, enabling accurate decision-making for clinical interventions. This retrospective study analysed the clinical data of 234 children with Mycoplasma pneumonia admitted to Cangnan Hospital of Wenzhou Medical University between March 2021 and October 2023. The data included general information, clinical symptoms, laboratory examination, and pathogenic profiles. The children were randomly divided into a training set (n = 164) and a validation set (n = 70) in a 7:3 ratio. Based on the efficacy of azithromycin sequential therapy, children in the training set were further divided into a poor efficacy group (n = 36) and a good efficacy group (n = 128). Independent risk factors for Mycoplasma pneumonia in the training set were identified using multiple logistic regression analysis. Furthermore, a column chart prediction model was constructed, and the model's performance was evaluated using receiver operating characteristic (ROC) curve analysis, followed by calibration curves. The predictive model was validated using an independent validation set, and decision curve analysis (DCA) assessed the model's clinical utility. In the training set, 36 cases (21.95%) showed poor therapeutic effects, while 24 cases (34.29%) in the validation set exhibited poor treatment response. There was no significant difference in clinical data between the two groups ( > 0.05). Univariate analysis showed significant differences ( < 0.05) across several factors, such as fever duration, cough severity, presence of pulmonary wet rales, white blood cell count, C-reactive protein (CRP) levels, Mycoplasma antibody titers, and Mycoplasma nucleic acid test findings among different treatment groups. Logistic regression analysis revealed prolonged fever duration, severe cough, presence of wet rales in the lungs, high white blood cell count, high CRP levels, high Mycoplasma antibody titers, and positive Mycoplasma nucleic acid test as independent risk factors of poor efficacy for azithromycin sequential treatment ( < 0.05). The Concordance index (C-index) of the column chart model was 0.804 in the training set and 0.861 in the validation set. The average absolute errors of the predicted and actual values were 0.129 and 0.081, respectively. The Hosmer-Lemeshow test results were χ = 10.288, = 0.245 for the training set and χ = 7.922, = 0.441 for the validation set, suggesting good model calibration. The ROC curve analysis revealed that the area under the ROC curve (AUC) for predicting the poor efficacy of azithromycin sequential therapy was 0.802 (95% confidence interval [CI]: 0.698-0.907) and 0.861 (95% CI: 0.704-1.000) for training and validation sets, respectively. Sensitivity and specificity were 0.655 and 0.907 in the training set and 0.898 and 0.952 in the validation set. Sensitivity analysis revealed that the model performed well across the decision subgroups, and the decision curve analysis indicated that the model demonstrated significant advantages when the threshold probability ranged between 0.1 and 0.98. This study is the first to construct a column chart prediction model using the clinical characteristics of Mycoplasma pneumonia in children, addressing the lack of prediction tools in this area. This model can offer a valuable reference for assessing the prognosis of azithromycin sequential treatment, helping clinicians develop more targeted and individualised treatment strategies.