Overestimating one's own health risks is associated with negative affect and decreased well-being. To identify psychosocial factors that reduce pessimistic risk estimates, the authors examined global perceived (primary and secondary) control as a predictor of health risk (hip fracture) estimates among 222 community-dwelling older adults. To determine whether characteristics of the health risk moderated the effects of perceived control on risk estimates, the authors manipulated risk level (low, high) and risk attribution (controllable, uncontrollable). The effects of perceived control differed as a function of risk attribution: Regardless of implied risk level, perceived primary control significantly predicted lower risk estimates in the controllable condition but not in the uncontrollable condition. In contrast, perceived secondary control significantly predicted lower risk estimates in the uncontrollable condition but not in the controllable condition, emphasizing its importance when direct influence is not feasible. The authors discuss implications for anticipating how older adults estimate their health risks.