Ovarian cancers, especially high-grade serous ovarian cancer (HGSOC), are one of the most lethal age-independent gynecologic malignancies. Although pathogenic microorganisms have been demonstrated to participate in the pathogenesis of multiple types of tumors, their potential roles in the development of ovarian cancer remain unclear. To gain an insight into the microbiome-associated pathogenesis of ovarian cancer and identify potential diagnostic biomarkers, we applied different techniques to analyse the microbiome and serum metabolome of different resources. We found that the vaginal microbiota in ovarian cancer mouse models was under dysbiosis, with altered metabolite configurations that may result from amino acid or lysophospholipid metabolic processes. Local therapeutic intervention with a broad spectrum of antibiotics was effective in reversing microbiota dysbiosis and suppressing carcinogenic progression. As the ovary is situated deeply in the pelvis, it is difficult to directly monitor the ovarian microbial community. Our findings provide alternative options for utilizing the vaginal bacteria as noninvasive biomarkers, such as Burkholderia (area under the curve = 0.8843, 95% confidence interval: 0.743-1.000), which supplement the current invasive diagnostic methods for monitoring ovarian cancer progression and contribute to the development of advanced microbe-based diagnosis and adjuvant therapies.© 2023 Wiley Periodicals LLC.