We aimed to assess the association between meditation practice and cognitive function over time among middle-aged and older adults. We included Health and Retirement Study (HRS) participants assessed for meditation practice in the year 2000 as part of the HRS alternative medicine module (n = 1,160) and were followed up for outcomes over 2000-2016 period. We examined the association between meditation ≥ twice a week vs none/less frequent practice and changes in the outcomes of recall, global cognitive function, and quantitative reasoning using generalized linear regression models. Stratified analyses among persons with/without self-reported baseline depressive symptoms were conducted to assess the link between meditation and cognitive outcomes. Among our full study sample, meditation ≥ twice a week was not significantly associated with total recall [; 95% CI: -0.97, 0.57; p = 0.61], global cognitive function [; 95% CI: -1.01, 1.12; p = 0.92], and quantitative reasoning [; 95% CI: -31.27, 8.32; p = 0.26]. However, among those who did not have self-reported depressive symptoms at baseline, meditation ≥ twice a week was associated with improvement in cognitive outcomes such as total recall [; 95% CI: 0.03, 0.18; p = 0.01] and global cognitive function [; 95% CI: 0.05, 0.40; p = 0.01] over time. Frequent meditation practice might have a protective effect on cognitive outcomes over time, but this protection could be limited to those without self-reported baseline depressive symptoms. Future studies could incorporate more precise meditation practice assessment, investigate the effect of meditation on cognitive outcomes over time, and include more rigorous study designs with randomized group assignment. This study is not preregistered.