Uncontrolled hemorrhage remains the leading cause of trauma-related death. The QuikClot is the most popular hemostatic device; however, it also shows severe security hidden dangers. This study aimed to evaluate the efficacy and safety of the self-developed Zeolite Dressing in a rabbit model of hemorrhagic groin injury. A complex hemorrhagic groin injury model was prepared in the right femoral artery of 96 rabbits by lacerating at a half depth and a subsequent 30-second spontaneous hemorrhage, and security was evaluated at the appointed time (1 and 3 weeks after operation). Experimental rabbits were randomly divided into four groups (12 males and 12 females in each group), including sham-operated control group without vascular injury (Con), standard gauze dressing (SD) group, QuikClot (QC) group and Zeolite Dressing (ZD) group. Vascular suture operation and postoperative debridement were thoroughly performed after hemostasis. The results showed that bleeding times of QC and ZD groups were significantly shorter than that of SD group (p<0.001), and there was non-significant difference between QC and ZD groups. Similarly, total blood loss was significantly reduced after the intervention of QC or ZD compared with SD group (p<0.001). No obvious toxic reaction was recorded 3 weeks after the treatment in both ZD and QC groups. A high wound temperature (maximum temperature was 96.6±27.9°C) and the “cooked meat like” pathological changes in the skin and muscle were only found in QC group. In conclusion, Zeolite Dressing exhibited a commensurate power to stop bleeding with higher safety compared with QuikClot, indicating that it is a practicable alternative for treating severe hemorrhage.