Introduction: Circulating androgen levels in men decline with advancing age and have been linked to diabetes and atherosclerotic cardiovascular disease (ASCVD). A similar relationship has been reported for low total testosterone (T) and incident heart failure (HF), but whether this differs for free T or the more potent androgen, dihydrotestosterone (DHT), remains unknown. Methods: We hypothesized that total and free T are inversely related and SHBG positively related to incident HF in older men, whereas, based on earlier data on ASCVD in CHS, total DHT bears a U-shaped relationship with this outcome. In a male sample from CHS without prevalent ASCVD or HF, serum T and DHT concentration were measured by liquid chromatography tandem mass spectrometry, and sex hormone-binding globulin (SHBG) by immunoassay. Free T was calculated from total T, SHBG and albumin. We used Cox regression to estimate relative risks of HF, adjusting for age, race, BMI, smoking, alcohol, physical activity, systolic BP, antihypertensive use, total and HDL cholesterol, and eGFR. Results: In 1061 men (age 76±5) followed for a median of 9.6 years, there were 368 HF events. Restricted cubic splines showed no meaningful departure from linearity, and relationships were expressed per SD decrement in androgen levels. After adjustment, lower free T was significantly associated with higher risk of HF (HR 1.14 [95% CI 1.01, 1.28]), p=0.03). Risk estimates for total T (HR 1.12 [0.99, 1.26], p=0.07), DHT (HR 1.10 [0.97, 1.24], p=0.14) and SHBG (HR 1.07 [0.95, 1.21], p=0.25) were directionally similar, but not statistically significant. The association of lower free T and HF was not altered after further adjustment for diabetes or atrial fibrillation. In a sensitivity analysis excluding participants with history of prostate cancer, findings were similar. Conclusion: Total T, free T, DHT and SHBG were inversely associated with incident HF, but only the association of free T was statistically significant. These findings suggest a contribution of androgen deficiency to HF incidence among men late in life, the segment of the population at highest risk of both conditions. Additional research is needed to determine whether androgen replacement therapy is an effective strategy to lower HF risk in older men.