Abstract: OBJECTIVE To explore the degree of cervical lesions and distribution of vaginal flora in the patients with high-risk human papillomavirus (HR-HPV) infection and analyze the association between them. METHODS Totally 160 patients who were tested positive for HR-HPV in the First Affiliated Hospital of Changjiang University from Jan 2021 to De 2022 were assigned as the positive group and were divided into the low-grade squamous intraepithelial lesions (LSIL) group with 75 cases (CIN grade 1), the high-grade squamous intraepithelial lesions (HSIL) group with 60 cases (CIN grade 2-3) and the cervical cancer group with 25 cases (diagnosed by cervical biopsy) according to the degree of cervical lesions. Meanwhile, 180 patients who were tested negative for HR-HPV were chosen as the negative group. The distribution of vaginal flora was observed and compared between the positive group and the negative group and among the patients with different degree of cervical lesions, the risk factors for the HR-HPV infection were analyzed, and a prediction model was established to analyze the predictive values. RESULTS The isolation rates of Chlamydia trachomatis, Ureaplasma urealyticum, bacteria and Candida of the positive group were higher than those of the negative group (P<0.05). The isolation rates of C.trachomatis, U.urealyticum, bacteria and Candida of the cervical cancer group were higher than those of the HSIL group and the LSIL group (P<0.05). U.urealyticum infection and bacterial vaginitis were the risk factors for the HR-HPV infection (P<0.05). The prediction model was as follows: logit(P)=-10.331+U.urealyticum infection×0.973+bactieral vaginitis×0.795. When logit(P) was more than 9.85, the area under curve (AUC) was 0.869,95%CI was 0.828-0.903, the sensitivity was 78.12%, and the specificity was 77.78%. CONCLUSION The imbalance of vaginal flora, the U.urealyticum infection and bacterial vaginitis in particularly, is closely associated with the HR-HPV infection and degree of cervical lesions. The prediction model can accurately predict the occurrence of HR-HPV infection, and it has high clinical value.