To assess the bidirectional association between chronic pain and depressive symptoms among middle-aged and older adults from two prospective cohort studies. We used prospective data (12y of follow-up) from the English Longitudinal Study of Ageing (n = 9149, 5018 women, 65.0 ± 10.2y) and the Health and Retirement Study (n = 16,883, 9810 women, 66.9 ± 10.3y), including data from seven waves of each cohort between 2006 and 2018/2019. Depressive symptoms were assessed using the Centre of Epidemiological Studies Depression scale, while chronic pain was estimated using questions about the frequency of being troubled with pain. We used random-intercept cross-lagged panel models to assess the bidirectional association between pain and depressive symptoms, adjusting for potential confounders. There was a cross-lagged effect of chronic pain on depressive symptoms (ELSA: β: 0.038; 95%CI: 0.011–0.066. Standardized coefficient (B): 0.021. SHARE: β: 0.044; 95%CI: 0.023–0.065. B: 0.023–0.024) as well as depressive symptoms on pain (ELSA: β: 0.010; 95%CI: 0.002–0.018. B: 0.017–0.019. SHARE: 0.011; 95%CI: 0.005–0.017. B: 0.020–0.021). Moreover, there were auto-regressive effects of both chronic pain (ELSA: β: 0.149; 95%CI: 0.128–0.171. SHARE: β: 0.129; 95%CI: 0.112–0.145) and depressive symptoms (ELSA: β: 0.149; 95%CI: 0.130–0.168. SHARE: β: 0.169; 95%CI: 0.154–0.184). We identified a modest bidirectional association between depressive symptoms and chronic pain, using two large prospective ageing cohorts. •The stable trait variances of both chronic pain and depressive symptoms were correlated in both cohorts.•There is a relatively low bidirectional relationship between chronic pain and depressive symptoms.•Strategies to address both conditions are needed in middle-aged and older adults.