Advocating continued health into old age, so called successful aging, is a growing public health goal. However, the development of tools to measure aging is limited by the lack of appropriate outcome measures, and operational definitions of successful aging. Using exploratory factor analysis, we attempted to identify distinguishable health domains with representative variables of physical function, cognitive status, social interactions, psychological status, blood biomarkers, disease history, and socioeconomic status from the InCHIANTI study. We then used logistic and mixed effect regression models to determine whether the resulting domains predicted outcomes of successful aging over a nine-year follow-up. A four-domain health model was identified: neuro-sensory function, muscle function, cardio-metabolic function and adiposity. After adjustment for age and gender, all domains contributed to the prediction of walking speed (R2=0.73). Only the muscle function domain predicted dependency (R2=0.50). None of the domains were a strong, significant predictor of self-rated health (R2=0.18) and emotional vitality (R2=0.23). Cross-sectional findings were essentially replicated in the longitudinal analysis extended to nine-year follow-up. Our results suggest a multi-domain health model can predict objective but not subjective measures of successful aging.