Background: Early identification of mood disorder post-stroke (MDPS) or its determinants could improve stroke outcomes. However, the natural history, prevalence and determinants of MDPS within the first weeks post-stroke require further investigation., Methods: Consecutive hospitalised stroke survivors were assessed within 2-5 days of stroke, and at 1 and 3 months post-stroke. Baseline data included demographics, co-morbidities, stroke subtype, pre-stroke disability and cognition. At baseline, 1- and 3-month interviews physical impairment, disability, cognition and social support were assessed. MDPS was defined as a score of >8 on the depression subscale of the Hospital Anxiety Depression Scale. Factors independently associated with MDPS at each time-point were determined using regression analyses., Results: One hundred and twenty-five subjects were included. The prevalence of MDPS within 5 days and at 1 and 3 months post-stroke was 5%, 16% and 21% respectively. The independent determinants for MDPS at 1 month were disability, social support and change in impairment score between initial and 1-month assessments; and at 3 months were disability, social support and institutionalisation. Individuals moved in and out of the subset of depressed patients over time. MDPS was independently associated with mortality at 3 months post-stroke., Conclusion: Mood disorder post-stroke increases in prevalence over the initial weeks post-stroke despite an improvement in disability, and is associated with mortality. Patients with MDPS at 1 month were not necessarily affected at 3 months and vice versa, indicating the dynamic nature of MDPS in the early stages.