PURPOSE: To examine the relationship between biological aging metrics and cardiovascular health, as well as the mediating effect of sleep duration. METHOD: We applied the recommended sampling weights to adjust for the complex survey design of NHANES. Using NHANES data, we first employed restricted cubic spline (RCS) and logistic regression models to explore the cross-sectional associations between biological aging metrics, defined by the Klemera-Doubal method biological age (KDM-BA), phenotypic age (PA), homeostatic dysregulation (HD), and allostatic load (AL), and the prevalence of cardiovascular diseases (CVD) and its subtypes. We then used Cox regression, Kaplan-Meier curves, and RCS models to assess the prospective associations between biological aging metrics and all-cause as well as CVD mortality. Further, ROC and DCA models were used to assess the predicting ability of 4 biological aging metrics to cardiovascular health. RESULT: This study included 7,704 participants. We found that biological aging metrics were strongly linked to the prevalence of CVD and its subtypes, as well as to all-cause and CVD mortality. Sleep duration appeared to moderate these associations. Among the four biological aging metrics, PA was the most effective predictor of CVD prevalence and its subtypes, though none of the metrics accurately predicted mortality. CONCLUSION: Biological aging metrics were significantly associated with cardiovascular health, while sleep duration may attenuate this relationship. Clinically, PA can be a potential predictor of cardiovascular health.