Abstract Context Emerging evidence has related the gut microbiome and circulating metabolites to human obesity. Gut microbiota is responsible for several metabolic functions, and altered plasma metabolome might reflect differences in the gut microbiome. Objective To identify a plasma metabolite profile associated with body mass index (BMI) in a general population and investigate whether such metabolite profile is associated with distinct composition of the gut microbiota. Design Targeted profiling of 48 plasma metabolites was performed in a population of 920 Swedish adults (mean age, 39 years; 53% women) from the ongoing Malmö Offspring Study using targeted liquid chromatography–mass spectrometry. Gut microbiota was analyzed by sequencing the 16S ribosomal RNA gene (V1-V3 region) in fecal samples of 674 study participants. Results BMI was associated with 19 metabolites (P < 0.001 for all), of which glutamate provided the strongest direct association (P = 5.2e-53). By orthogonal partial least squares regression, a metabolite principal component predictive of BMI was constructed (PCBMI). In addition to glutamate, PCBMI was dominated by branched-chain amino acids (BCAAs) and related metabolites. Four gut microbiota genera (Blautia, Dorea, Ruminococcus, and SHA-98) were associated with both BMI and PCBMI (P < 8.0e-4 for all). When simultaneously regressing PCBMI and metabolite-associated gut bacteria against BMI, only PCBMI remained statistically significant. Conclusions We discovered associations between four gut microbiota genera (Blautia, Dorea, Ruminococcus, and SHA-98) and BMI-predictive plasma metabolites, including glutamate and BCAAs. Thus, these metabolites could be mediators between gut microbiota and obesity, pointing to potential future opportunities for targeting the gut microbiota in prevention of obesity. We identified metabolites associated with BMI. The BMI-related metabolites were connected to four microbiota genera, suggesting a mediating role of the metabolites in the microbiota–obesity relationship.