To develop and validate a nomogram for predicting recurrent lumbar disk herniation (LDH) within 2 years after percutaneous endoscopic lumbar discectomy. Information on patients' LDH was collected from 1 medical center between January 2015 and September 2020. The LASSO (least absolute shrinkage and selection operator) method was applied to select the most significant risk factors. A multivariate logistic regression analysis was used to develop a predictive model incorporating the possible factors selected by the LASSO regression model. The discriminant, corrected, and clinically useful prediction models were evaluated using consistency index (C-index), receiver operating characteristic curve, calibration curves, and decision curve analysis. Internal validation of clinical predictive power was also assessed by bootstrap validation. A total of 690 patients with LDH were included in this study. Sixty-three patients experienced recurrence within 2 years whereas 627 experienced no recurrence. The nomogram predictors included age, body mass index, Modic change, Pfirrmann grade, and sagittal range of motion. The model had good discrimination power, with a reliable C-index of 0.868 (95% confidence interval, 0.822-0.913) and interval validation confirmed a higher C-index value of 0.846. The area under the receiver operating characteristic curve was 0.868, indicating a good predictive value. The decision curve analysis indicated that it was clinically feasible to use the predictive recurrence nomogram model. We developed and validated a new accurate and effective nomogram for predicting recurrent LDH within 2 years after percutaneous endoscopic lumbar discectomy. Age, body mass index, Modic change, Pfirrmann grade, and sagittal range of motion were significant features for predicting rLDH.