During the last decade, sow mortality due to pelvic organ prolapse (POP) has increased. To better understand the biology associated with POP, sows were phenotypically assessed and assigned a perineal score (PS) based on presumed POP risk and categorized as PS1 (low), PS2 (moderate), or PS3 (high). The study objective was to identify changes in sow vaginal microbiota that may be associated with POP. The hypothesis is that vaginal microbiota differs between sows with variable risk for POP, and changes in microbiota during late gestation exist between sows with differing risk. Of the 2864 sows scored during gestation week 15, 1.0, 2.7, and 23.4% of PS1, PS2, and PS3 sows, respectively, subsequently experienced POP. Vaginal swabs subjected to 16S rRNA gene sequencing revealed differences in community composition (Bray–Curtis; P < 0.05) and individual operational taxonomic unit (OTU) comparisons between vaginal microbiota of PS1 and PS3 sows at gestation week 15. Further, differences ( P < 0.05) in community composition and OTUs ( Q < 0.05) were observed in PS3 sows that either did or did not subsequently experience POP. Differences in community structure (alpha diversity measurements; P < 0.05), composition ( P < 0.05), and OTUs ( Q < 0.05) were observed in gestation week 12 sows scored PS1 compared to week 15 sows scored PS1 or PS3, suggesting that sow vaginal microbiota shifts during late gestation differently as POP risk changes. Collectively, these data demonstrate that sows with greater POP risk have unique vaginal microflora, for which a better understanding could aid in the development of mitigation strategies.