ObjectiveTo examine depressed affect, somatic complaints, and positive affect as longitudinal predictors of fluid, crystallized and global cognitive performance in the Kuakini Honolulu-Asia Aging Study (HAAS), a large prospective cohort study of Japanese-American men. MethodsWe assessed 3,088 dementia-free Kuakini-HAAS participants aged 71–93 (77.1 ± 4.2) years at baseline (1991–1993). Depressive symptoms were evaluated by the Center for Epidemiologic Studies Depression (CES-D) Scale. Baseline CES-D depression subscales (depressed and positive affects; somatic complaints) were computed. The Cognitive Abilities Screening Instrument (CASI) measured cognitive performance on a 100-point scale; fluid and crystallized cognitive abilities were derived from CASI factor analysis. Cognition was also evaluated at 4 follow-up examinations over a 20-year period. Multiple regression assessed baseline CES-D subscales as predictors of cognitive change. The baseline covariates analyzed were CASI, age, education, prevalent stroke, APOE ε4 presence, and the longevity-associated FOXO3 genotype. ResultsCross-sectionally, baseline CES-D subscales were related to cognitive measures; e.g., higher depressed affect was associated with lower crystallized ability (β = −0.058, p ≤ 0.01), and somatic complaints were linked to poorer fluid ability (β = −0.045, p ≤ 0.05) and to worse global cognitive function as measured by total CASI score (β = −0.038, p ≤ 0.05). However, depression subscales did not significantly or consistently predict fluid ability, crystallized ability, or global cognitive performance over time. ConclusionPsychological and physical well-being were associated with contemporaneous but not subsequent cognitive functioning. Assessment of depressive symptoms may identify individuals who are likely to benefit from interventions to improve mood and somatic health and thereby maintain or enhance cognition.