Associations of serum calcium (S-Ca) and 25-hydroxyvitamin D (S-25(OH)D) concentrations with longevity, cardiovascular disease, and cancer are not clear. We conducted a Mendelian randomization study to examine the associations of S-Ca and S-25(OH)D with longevity and risk of cardiovascular disease and cancer. The primary genetic instruments for S-Ca and S-25(OH)D were obtained from genome-wide association meta-analyses that included 61,054 individuals for S-Ca and up to 79,366 individuals for S-25(OH)D. Genetic variants associated with S-Ca and S-25(OH)D in the UK Biobank were used as confirmatory instruments. We obtained summary-level data for associations of these instruments with individual survival later than the 90 versus at most the 60 percentile of expected age at death from a genome-wide association meta-analysis including 11,262 cases and 25,483 controls, and with parental longevity (both parents in top 10% percentile) from the UK Biobank including 7,182 cases and 79,767 controls. Data for cardiovascular disease (111,108 cases and 107,684 controls) and cancer (38,036 cases and 180,756 controls) were obtained from the FinnGen consortium. A one standard deviation increase in genetically-predicted S-Ca concentration was associated with lower odds of longevity (odds ratio, 0.72; 95% CI, 0.55-0.95) and increased risk of cardiovascular disease (odds ratio, 1.11; 95% CI, 1.03-1.20). The associations were consistent in confirmatory analyses. There was no evidence supporting an association between genetically-predicted S-Ca and cancer, and no associations of genetically-predicted S-25(OH)D with the studied outcomes. Lifelong higher levels of S-Ca but not S-25(OH)D may shorten life expectancy and increase the risk of cardiovascular disease.