Objective: To evaluate the effectiveness of super-selective renal artery embolization in treatment of post-percutaneous nephrolithotomy bleeding, and to analyse the causes of failure embolization. Methods: In the study, 65 post-percutaneous nephrolithotomy patients with severe renal bleeding and hemodynamic instability were treated by super-selective renal artery embolization. First of all, we performed selective renal arteriography. After clarifying the location of the bleeding, superselective intubation of the injured vessel with a microcatheter was carried out. Then the injured vessel was embolized with Tornado micro-coil. When complete embolization was not achieved with micro-coil, a small amount of gelatin sponge particles were added. If there was no positive finding of the beginning selective renal arteriography, the following measures could be taken to prevent missing lesions: (1) Abdominal aorta angiography was performed to determine whether there were anatomical variations, such as accessory renal arteries or multiple renal arteries; (2) Ultra-selective intubation angiography next to the nephrostomy tube path was performed; (3) Renal arteriography was repeated; (4) Renal arteriography after removing the nephrostomy tube while retaining the puncture channel. We evaluated the different angiographic findings and analysed the causes of embolization failure. Results: Bleeding was successfully controled in 60 patients (62 kidneys) whose renal arteriography was postive. Positive findings included: pseudoaneurysm formation, patchy contrast extravasation, pseudoaneurysm combined with arteriovenous fistula, contrast agent entering the collection system, extravascular perinephric leakage of contrast. After first embolization, bleeding was controled in 53 patients (55 kidneys). The success rate after the first and second embolization was 88.7% and 96.7% respectively. The second session was required because of failure to demonstrate bleeding arteries during the first session (4 patients, 57.1%) and recurrent hemorrhage of the embolized injured arteries (2 patients, 28.6%). In 5 patients with no positive findings, after conservative treatment, hematuria disappeared. All the patients were followed up for 3, 6, and 12 months after embolization, and no hematuria occurred again, and no sustained and serious renal insufficiency. Conclusion: Super-selective renal artery embolization is an effective treatment for post percutaneous nephrolithotomy bleeding. The main cause of failure is omitting of injured arteries during renal arteriography. Renal artery branch injury has various manifestations. Attention should paid to the anatomical variation of the renal artery, and patient and meticulous superselective intubation angiography is the key to avoiding missing the lesion and improving the success rate of embolization.