Objective To identify prognostic factors in children with severe Mycoplasma pneumoniae pneumonia （SMPP） using lung ultrasound and clinical indicators， and to construct and evaluate a nomogram-based predictive model. Methods A total of 527 children with SMPP treated at Bijie Hospital of Zhejiang Provincial People's Hospital between January 2021 and December 2024 were enrolled and randomly allocated into a training set （n = 368） and a testing set （n = 159） in a 7∶3 ratio. The training set was categorized into good prognosis group （n = 315） and poor prognosis group （n = 53） based on the children's post-treatment conditions. Univariate and multivariate logistic regression analyses were performed to assess the correlation between patients' lung ultrasound and clinical indicators； receiver operating characteristic （ROC）， calibration curve， and decision curve analysis （DCA） were used to evaluate the constructed nomogram predictive model. Results The poor prognosis group had a longer duration of fever at admission， higher proportion of high fever， higher white blood cell （WBC） count， higher absolute neutrophil count， D-dimer， higher lactate dehydrogenase （LDH）， larger consolidation size to body surface area ratio （consolidation size/BSA）， liquid-bronchogram sign， and higher proportion of pleural effusion compared to the good prognosis group （P < 0.05）， while only the air-bronchogram sign was lower in the poor prognosis group （P < 0.05）； the duration of fever during treatment， absolute neutrophil count， DD dimer， LDH levels， and pleural effusion were independent risk factors for poor prognosis in SMPP children. The area under the curve （AUC） for the training set was 0.926 （95%CI： 0.870 - 0.982）， and for the test set was 0.957 （95%CI： 0.882 - 1.000）. Calibration curve and DCA curve assessment of the nomogram predictive model showed it had high consistency and clinical application value. Conclusion The duration of fever during treatment， absolute neutrophil count， DD dimer， LDH levels， and pleural effusion are independent risk factors for poor prognosis in children with SMPP， and the constructed nomogram predictive model has good predictive performance.