Bacterial vaginosis (BV) is a syndrome of the female reproductive tract associated with adverse reproductive outcomes and characterized by a shift from a Lactobacillus (LB)-dominant vaginal microbiota to a polymicrobial, anaerobic microbiota, consistently colonized by strains of Gardnerella vaginalis (Gv). The first-line treatment for BV is metronidazole (MNZ); however, treatment failure and recurrence rates remain high. To gain insight into complex interactions between target species (Gv) and non-target Lactobacillus species (Lactobacillus iners (Li)) with MNZ and understand their respective roles in efficacy, we developed an ordinary differential equation model that predicts bacterial growth as a function of drug uptake, metabolism, proliferation, and MNZ sensitivity. Model findings revealed a critical factor in MNZ efficacy may be Li sequestration of MNZ, and that efficacy decreases when the relative abundance of Li is higher pre-treatment. These results were validated in Gv and Li co-cultures (p < 0.001), and in two clinical cohorts, finding women with recurrent BV had significantly lower pre-treatment levels of BV-associated bacteria relative to Lactobacillus spp. (p = 0.0366; p = 0.0484). Overall, model results support a mechanism where non-target Lactobacillus species sequester MNZ from BV-associated target species, such as Gv, promoting BV recurrence by reducing MNZ bioavailability.