To evaluate risk factors of severe adverse events after percutaneous nephrolithotomy with an emphasis on operation time, and to develop a nomogram for predicting them. This was an observational retrospective study including 1511 patients who underwent percutaneous nephrolithotomy in 332 hospitals identified from the Japanese Diagnosis Procedure Combination database between 2007 and 2010. Severe adverse events were defined as follows: (i) in-hospital mortality; (ii) postoperative medications including catecholamine, gamma-globulin products, protease inhibitors and medications for disseminated intravascular coagulation; and (iii) postoperative interventions including central vein catheterization, dialysis and mechanical cardiopulmonary support. Univariate and multivariate logistic regression analyses were carried out for the occurrence of severe adverse events, and a nomogram was generated from this model. Overall, 126 severe adverse events (8.34%) were identified. In the multivariate model, a linear trend between severe adverse events and operation time was observed (OR 4.72 for 120-179 min to 17.95 for ≥300 min compared with ≤119 min; each P < 0.05) after adjustment for sex, age, Charlson Comorbidity Index and type of admission. Female sex and emergency admission were also significant risk factors (OR 1.92 and 2.04, respectively), and hospital volume did not reach statistical significance. The nomogram based on these results was well fitted to predict a probability between 0.05 and 0.40 (concordance index 0.696). Longer operation time is a significant and independent risk factor for severe adverse events after percutaneous nephrolithotomy. Our nomogram can be an effective tool for predicting postoperative complications.