GOAL: We utilize group-based trajectory modeling (GBTM) to delineate depressive symptom trajectories among stroke survivor-caregiver dyads, to identify predictors of the delineated trajectories, and to assess the influence of time-varying covariates (stroke survivor depressive symptoms and functional disability, caregiver depressive symptoms, and foreign domestic worker [FDW] assistance) on the level of the depressive symptom trajectories. METHODS: Data on 172 stroke survivor-caregiver dyads in Singapore, for whom depressive symptoms were assessed thrice (baseline/3 months/6 months), were utilized. GBTM was applied to delineate depressive symptom trajectories, and to identify their predictors and time-varying covariates. FINDINGS: Three stroke survivor depressive symptom trajectories (low and decreasing [47.6%], low and increasing [43.1%], and high and increasing [9.3%]) and 2 caregiver depressive symptom trajectories (low and stable [71.5%] and high and decreasing [28.5%]) were delineated. Caregivers with chronic diseases were more likely (odds ratio [95% confidence interval]: 8.09[2.04-32.07]) and those caring for older stroke survivors (0.94[0.90-0.98]) were less likely to follow the high and decreasing than the low and stable depressive symptom trajectory. An increase in stroke survivor functional disability and caregiver depressive symptoms led to a rise (~worsening) in stroke survivor depressive symptom trajectories. Whereas an increase in stroke survivor depressive symptoms led to a rise in caregiver depressive symptom trajectories, FDW assistance led to a decline (~improvement). CONCLUSION: Care professionals should be mindful of heterogeneity in depressive symptom patterns over time among stroke survivor-caregiver dyads. Reciprocal association of depressive symptoms in the stroke survivor-caregiver dyad suggests that addressing mood problems in 1 member may benefit the other member, and calls for dyadic mental health interventions.