To test the hypothesis that frontal lobe functions mediate effects of religiosity on health, verbal and design fluency tasks, religiosity scales and self-report measures of health were administered to 80 community dwelling older persons. Low to moderate correlations were obtained between measures of fluency performance, intrinsic religiosity, and self-reported number of illnesses. Participants with both high fluency scores and high prayer (frequency) scores evidenced significantly better health scores than subjects with only high fluency performance or high prayer scores. Participants with both low fluency and low religiosity evidenced the poorest health scores. These data support the view that neuropsychologic measures may interact with private religious practices in producing better self-reported health in older people.