Objective To comprehensively analyze the clinical characteristics and risk factors of children with refractory mycoplasma pneumonia （RMPP）， precisely identify the joint predictors in these children， and construct a prediction model. This aims to offer a scientific foundation for the early identification of RMPP and the formulation of accurate treatment and medication strategies. Methods A retrospective analysis was performed on the clinical data of 282 children diagnosed with mycoplasma pneumoniae pneumonia （MPP） who were admitted to the Pediatric Department of the Seventh Medical Center of the Chinese People's Liberation Army General Hospital between August 1， 2023， and February 29， 2024. Among these children， 119 with RMPP were classified into the RMPP group， while the remaining 163 with general MPP （GMPP） were assigned to the GMPP group. The clinical data of both groups， encompassing age， gender， duration of fever， symptoms， laboratory test indices， chest imaging data， complications， etc.， were compared. A logistic probability model （LogP model） for joint application was constructed. The discriminatory ability of the model was evaluated using the area under the receiver operating characteristic （ROC） curve， and the calibration of the model was assessed by means of a calibration curve. Results In comparison with the GMPP group， children in the RMPP group exhibited a significantly longer duration of fever （P = 0.002）. Moreover， they had a higher incidence of complications， including myocardial damage and coagulation dysfunction （P < 0.05）. Regarding inflammatory markers， the levels of C-reactive protein （CRP）， procalcitonin （PCT）， and lactate dehydrogenase （LDH） were notably elevated in the RMPP group （P < 0.05）， whereas the level of albumin （Alb） was lower （P = 0.001）. In the RMPP group， the levels of interleukin-2 （IL-2）， interleukin-5 （IL-5）， interleukin-8 （IL-8）， interleukin-1β （IL-1β）， and D-Dimer were increased， while the levels of interleukin-6 （IL-6） and interleukin-17 （IL-17） were decreased （P < 0.05）. Chest computed tomography （CT） scans revealed a higher proportion of lung consolidation， pleural effusion， and atelectasis in the RMPP group （P < 0.05）. Multivariate logistic regression analysis demonstrated that CRP， total bilirubin （T-BIL）， LDH， IL-17， and prothrombin time （PT） were independent risk factors for RMPP （P < 0.05）. The receiver operating characteristic （ROC） predictive model established based on these factors had an area under the curve （AUC） of 0.787 （95% confidence interval ［CI］： 0.693 ~ 0.880）， with a cutoff value of 0.421， a sensitivity of 0.786， and a specificity of 0.660. The calibration curve indicated that the predicted probability matched well with the reference probability， and there was no statistical difference in the results of the Hosmer-Lemeshow test （P > 0.05）. Conclusions The clinical features of children with RMPP are predominantly characterized by prolonged fever， moderate lung lesions， other organ injuries， and high inflammatory markers. CRP， T-BIL， LDH， IL-17， and PT can act as independent risk factors for RMPP.