PURPOSE: Valuation of life (VOL) represents a construct capturing active attachment to life put forward by M. P. Lawton (e.g., 1999). As old and very old individuals may differ in terms of endorsement and with respect to what makes a life worth living, the present study investigated whether mean levels and the explanatory value of sociodemographic, social, and health predictors for VOL differ between young-old and old-old individuals. DESIGN AND METHODS: We presented a sample of 356 community-dwelling individuals aged 65 to 94 years with Lawton's Positive Valuation of Life Scale and established measures to assess predictors of VOL. RESULTS: Mean levels of VOL decreased from the third to the fourth age. Zero-order correlations showed significant relations between sociodemographic (i.e., age, gender, marital status, education), social (i.e., social contacts, phone calls, volunteering, contact with youth), and health (i.e., mobility, vision, hearing, activity restrictions, activity of daily living [ADL], instrumental ADL [IADL]) indicators. Regression analyses for the domain-specific predictors reduced the number of significant predictors to age, education, grandchildren, vision, and IADLs. When combining all sets of predictors, health explained twice as much variance in VOL compared to social indicators; sociodemographic indicators including age made no independent contribution. Separate analysis for young-old and old-old participants revealed age-differential prediction patterns. For the young-old, the role of health factors was especially strong, whereas specific social factors became more important in the old-old group. IMPLICATIONS: Age-differential predictive values of the resources seem to indicate positive adaptation to aging. Taking into account such prediction patterns may help to design specific interventions for young-old and old-old individuals.