It is well-established that depression is common after stroke. Much less is known about the longitudinal course of depressive symptoms in stroke survivors. Data for this study came from the DESERVE trial of a skills-based intervention to reduce vascular risk in mild/moderate stroke/TIA patients. Depressive symptoms were assessed at baseline (pre-discharge), 6 months and 1 year post-discharge using the Center for Epidemiologic Studies - Depression (CES-D) scale. Discrete mixture models identified distinct trajectories of depressive symptom severity over time and assigned patients to trajectory groups. We ran Poisson models using linear and quadratic parameters and chose the best-fitting model based on Bayesian information criterion and interpretability. Associations between baseline characteristics and trajectory group membership were examined using ANOVA and bivariable multinomial logistic regression. So far, 285 patients completed all three study waves and ≥ 1 depression assessment. A four-group model was selected: 1) resistance - low depressive symptom level at all waves (50% of patients); 2) delayed - low baseline symptom level, increasing to clinically significant levels (CES-D > 16) by 6 months, remaining high at 1 year (18%); 3) recovery - clinically significant baseline symptom levels, dropping to low levels by 6 months, remaining low at 1 year (22%); 4) chronic - severe symptoms at all waves (10%). Hispanic patients had greater odds of delayed symptoms vs. resistance compared to non-Hispanics. Moderate stroke patients had greater odds of delayed and chronic symptoms vs. resistance compared to mild stroke patients. Patients with stroke history had greater odds of delayed symptoms and recovery vs. resistance compared to those without; those with psychiatric history had greater odds of recovery and chronic symptoms vs. resistance compared to those without. Those with chronic symptoms had significantly greater baseline disability compared to other groups. The course of post-stroke depressive symptoms is heterogeneous and associated with race/ethnicity and several clinical factors. Targeted interventions may be required to prevent the development of depression and address chronic symptoms that may interfere with stroke recovery.