Objectives: To determine the relationship between objectively measured physical activity (PA) and the gut microbiome among community-dwelling older men. Design: Cross-sectional study. Setting: Osteoporotic Fractures in Men (MrOS) cohort participants at Visit 4 (2014-16). Participants: Eligible men (n=373, mean age 84 y) included participants with 5-day activity assessment with at least 90% wear time and analyzed stool samples. Measurements: PA was measured with the SenseWear Pro3 Armband and stool samples analyzed for 16S v4 rRNA marker genes using Illumina MiSeq technology. Armband data together with sex, height, and weight were used to estimate total steps, total energy expenditure, and level of activity. 16S data was analyzed using standard UPARSE workflow. Shannon and Inverse Simpson indices were measures of (within-participant) α-diversity. Weighted and unweighted Unifrac were measures of (between-participant) β-diversity. We used linear regression analysis, principal coordinate analysis, zero-inflated Gaussian models to assess association between PA and α-diversity, β-diversity, and specific taxa, respectively, with adjustments for age, race, BMI, clinical center, library size, and number of chronic conditions. Results: PA was not associated with α-diversity. There was a slight association between PA and β-diversity (in particular the second principal coordinate). Compared to those who were less active, those who had higher step counts had higher relative abundance of Cetobacterium and lower relative abundance of taxa from the genera Coprobacillus, Adlercreutzia, Erysipelotrichaceae CC-115 after multivariable adjustment including age, BMI, and chronic conditions. There was no consistent pattern by phylum. Conclusion: There was a modest association between levels of PA and specific gut microbes among community-dwelling older men. The observed associations are consistent with the hypothesis that underlying health status and composition of the host microbiome are related.