Purpose: Poor sleep is associated with short-term dysregulation of mood and is a risk factor for major depressive disorder (MDD). This study examines whether objectively measured sleep in late adolescence prospectively predicts major depressive episode (MDE) onset in early adulthood as well as whether daily affect mediates this association. Methods: The present study draws on subjective and objective sleep data, ecological momentary assessment, and diagnostic data from the longitudinal Youth Emotion Project to examine whether: a) short sleep predicts dysregulated ecological momentary assessment-measured mood the next day; b) sleep predicts depressive episodes over the subsequent 5 years; and c) dysregulated daily moods mediate the associations between short sleep and later MDD. Fixed effects, logistic regression, and formal mediation analyses were employed. Results: Our results showed that nights with less sleep are followed by days with more negative affect; short sleep predicted MDEs over the subsequent 5 years (adjusting for prior MDD); and negative affect mediates the relationship between short sleep and later MDEs. Discussion: Overall, our findings show sleep to be an important risk factor and hence a promising point of intervention for improving mood and reducing the risk of future MDEs in adolescents and early adults.