Background: Respiratory syncytial virus (RSV) is a leading cause of severe lower respiratory tract infections in children. This study aimed to develop and validate a nomogram for early prediction of severe RSV infection by identifying independent risk factors. Methods: We retrospectively analyzed clinical data of children diagnosed with RSV infection, dividing them into severe and non-severe groups. Variable selection was performed using Elastic Net regression combined with clinical knowledge on five multiply imputed datasets. A multivariable logistic regression model was then constructed. Internal validation was conducted using 500 bootstrap resamples to obtain optimism-corrected AUC and calibration slope, which was applied as a shrinkage factor to adjust regression coefficients. Results: Of 2,595 children, 160 were in the severe group and 2,435 in the non-severe group. Five predictors were retained: neutrophil-to-lymphocyte ratio (NLR), age, winter onset, hypoxemia, and preterm birth. The pooled AUC was 0.847 (95% CI: 0.833–0.860), and the optimism-corrected AUC after bootstrap validation was 0.843. The calibration slope was 0.975, indicating low overfitting risk. Conclusions: A nomogram incorporating five easily available predictors was developed to predict severe RSV infection in children. Internal validation demonstrated good discrimination and low overfitting, providing a practical tool for early identification of high-risk patients.