The study compared the efficacy and safety of superselective and selective transcatheter arterial embolization (TAE) for non-variceal gastrointestinal bleeding (NVGIB) unresponsive to endoscopy. 116 patients with NVGIB were included, of which 88 with upper tract bleeding (NVUGIB) - 48 treated with superselective and 40 with selective embolization. Technical success was achieved in all patients. Clinical success was higher in the superselective embolization group (85.42%) compared to selective (70.00%), but the difference was not statistically significant. The rate of rebleeding within 3 days was significantly lower in the superselective embolization group (8.33% vs. 27.50%). Bleeding-related mortality was 6.25% in the superselective and 7.50% in the selective embolization group. There were no significant differences between methods in lower tract bleeding, but all 5 cases of intestinal ischemia occurred in this subgroup. The type of embolic material used significantly influenced rebleeding and the need for further treatment for lower tract bleeding.