Background Bacterial vaginosis (BV) is one of the most common infectious diseases among sexually active women and is associated with the increased acquisition of a variety of sexually transmitted diseases.This study aimed to compare the efficacy of a non-antibiotic sucrose gel against an antibiotic metronidazole gel for the treatment of BV.Methods A randomized, double-blinded, multi-center, parallel-group, placebo-controlled phase Ⅲ clinical trial was conducted at eight hospitals in China.A total of 560 subjects with clinically diagnosed BV were randomly assigned into three groups for vaginally receiving sucrose, metronidazole, and placebo gels, respectively, twice daily for five consecutive days.The efficacy of therapeutic cure, defined as an achievement of both microbiologic cure (a Nugent score of 3 or less) and clinical cure (a resolution of the clinical findings from the baseline visit), was evaluated at the 1st and 2nd test-of-cure (TOC) visits at 7-10 and 21-35 days after the start of treatment, respectively.Results Therapeutic cure rates for sucrose, metronidazole, and placebo gel groups were 83.13%, 71.30% and 0.92%,at the 1st TOC, and 61.04%, 66.67% and 7.34%, at the 2nd TOC, respectively.While there was no significant difference between the sucrose and metronidazole gel groups at the 2nd TOC (P=0.305), and sucrose gel was more effective than metronidazole gel at the 1st TOC (P=0.009).Conclusion These findings suggest that sucrose gel restores normal vaginal flora more rapidly than metronidazole gel and can be used as a novel treatment for BV.