BACKGROUND As a common pediatric respiratory disease, Mycoplasma pneumoniae pneumonia (MPP) accounts for 20 - 30% of acquired pneumonia in children, with a rising morbidity rate annually. We aimed to explore the correlations of different serological parameters with the severity and prognosis of MPP children. METHODS A total of 108 MPP children were divided into severe group (n = 58) and mild group (n = 50). The serum levels of triglycerides, HDL-C, LDL-C, erythrocyte sedimentation rate (ESR), D-dimer (DD), lactate dehydrogenase (LDH), interleukin-6 (IL-6), galectin-3 (Gal-3), homocysteine (Hcy), and procalcitonin (PCT) were measured. Receiver operating characteristic (ROC) curves were plotted to analyze their predictive values for poor prognosis. They were followed up for 6 months and assigned into good and poor prognosis groups. Multivariate logistic regression analysis was performed to explore the serological parameters affecting prognosis. A prediction model was established. RESULTS In acute and recovery phases, the levels of ESR, DD, LDH, IL-6, Gal-3, Hcy, and PCT in the severe group were significantly higher than those in the mild group (p < 0.05). Prediction combining various serological parameters had the highest value for poor prognosis. ESR, DD, LDH, IL-6, Gal-3, Hcy, and PCT levels were independent risk factors. The concordance index of the nomogram model established using these factors was 0.745 (95% CI: 0.684 - 0.830). The area under the ROC curve was 0.726 (95% CI: 0.701 - 0.815). The predicted probability of the model was consistent with the actual one, showing high accuracy. CONCLUSIONS Serum levels of ESR, DD, LDH, IL-6, Gal-3, Hcy, and PCT are closely correlated with the severity and prognosis of MPP children, which provide references for prognostic evaluation. Prediction combining these indices is more valuable than that using a single index.