BackgroundDepression has been found to be associated with cognitive decline, but whether longer depressive durations lead to more severe cognitive declines has not been investigated. We aimed to estimate the association between depressive duration and cognitive decline in middle-aged and older Americans based on a large-scale representative population study. MethodsWe included 27,886 participants from the Health and Retirement Study (HRS) in 2010–2018. Four datasets with 2-, 4-, 6-, and 8-year consecutive interviews were further derived which involving persistent depressed and persistent depression-free individuals. Multiple linear regressions were constructed to estimate the effects of each depressive duration on the decline in global cognition, memory and mental status. Meta-regressions were performed to test the linear trends and to explore the heterogeneity between sex, age and baseline cognitive function along with subgroup analyses. ResultsDepressive durations of 2, 4, 6, and 8 years were associated with reductions in global cognitive scores of 0.62 points (95% CI: 0.51–0.73), 0.77 points (95% CI: 0.60–0.94), 0.83 points (95% CI: 0.55–1.10), and 1.09 points (95% CI: 0.63–1.55), respectively, indicating a linear trend (P = 0.016). More pronounced associations were observed in middle-aged adults and females. Similar patterns were found in the associations between depressive duration and two subdomains, i.e., memory and mental health. LimitationsThis study is essentially a cross-sectional study and therefore cannot provide causal associations. ConclusionsLonger depressive durations were linearly related to more severe cognitive declines. Timely intervention for depression targeted middle-aged adults can more effectively alleviate cognition-related burdens.