Late life involves a variety of different challenges to well-being. This study extends and qualifies propositions drawn from the paradox of well-being in aging using 15-year longitudinal data on depressive symptoms from old and very old participants in the Australian Longitudinal Study of Ageing (Baseline N = 2,087; Mage = 78.69 years; range: 65–103 years; 49.40% women). We first examined age-related trajectories in depressive symptoms from young-old to oldest-old, taking into account (changes in) relevant correlates, pathology, and mortality; and, second, we investigated gender differences in these trajectories. Results revealed that age-related trajectories of depressive symptoms were predictive of mortality hazards. The unique predictive effects of both level of, and change in, depressive symptoms were independent of one another and held after taking into account education as well as changes in marital status, living arrangements, cognitive function, and illness burden. In addition, results indicated that depressive symptoms were elevated among participants suffering from arthritis, and increased with age more markedly in men than in women. In particular, the significant Age × Gender interaction indicated that the gender gap in depressive symptoms reduced from young-old to old-old and reversed in very old age when men showed more depressive symptoms than women. Qualifying the paradox of well-being in aging, findings demonstrated that depressive symptoms increased from young-old to oldest-old and suggest that age-, pathology-, and mortality-related changes should be examined in concert to advance our understanding of individual differences in depressive symptom trajectories in late life.