To report the preliminary result of empiric embolization for angiographycally-negative lower gastrointestinal bleeding (LGIB) by using the pharmaco-induced vasospasm technique with or without the adjunctive use of intra-arterial multi-detector computed tomography (MDCT). 23 LGIB patients with positive MDCT findings but negative angiographic results underwent empiric pharmaco-induced vasospasm therapy. The presumed bleeding artery was semi-selectively catheterized, and a segment of bowel was temporarily spasmed with bolus injection of epinephrine and immediately followed by 4-h’ vasopressin infusion. The rebleeding, primary and overall clinical success rates were reported. MDCT showed 19 bleeders in the SMA territory and 4 bleeders in the IMA territory. Early rebleeding was found in 6 patients (26.1%): 2 local rebleeding, 3 from new-foci bleeding and 1 uncertain. Of the 10 small bowel bleeding patients, only 1 out of the 7 who underwent intra-arterial MDCT showed rebleeding, whereas 2 out of the 3 without intra-arterial MDCT rebled. No patients exhibited procedure-related major complications, including bowel ischemia and cardiopulmonary distress. The overall clinical success rate was 91.3% (21/23) with a 30-day mortality rate of 26.1% (2 of the 6 patients had early rebleeding). Empiric pharmaco-induced vasospasm therapy, when localized with/without adjunctive intra-arterial MDCT, seems to be a safe and effective method to treat angiographically-negative LGIB patients.