OBJECTIVE: To determine predictors of child externalizing (behavioral) and internalizing (emotional) symptoms in a national population sample. METHODS: Data were collected in 3 biennial waves (2004, 2006, and 2008) from 2 cohorts in the Longitudinal Study of Australian Children, initially including 5107 children 0 to 1 year of age and 4983 children 4 to 5 years of age. The primary outcomes were child externalizing and internalizing symptoms. Relationships between potential risk factors and child mental health outcomes were described by using linear regression. RESULTS: In unadjusted analyses, children's mental health symptoms were predicted by a large number of risk factors. In multivariate models, early childhood factors (birth through 5 years) explained 30% and 18% of variations in externalizing and internalizing symptoms, respectively, at 4 to 5 years of age. Middle childhood (5-9 years of age) factors explained 20% and 23% of variations in externalizing and internalizing symptoms, respectively, at 8 to 9 years of age. Harsh discipline was a strong consistent predictor of externalizing symptoms in both age groups, whereas poorer child physical health, maternal emotional distress, harsh discipline, and overinvolved/protective parenting (younger cohort only) predicted internalizing symptoms consistently. CONCLUSIONS: National data on predictors of child mental health symptoms highlighted a small number of significant risk factors, situated in the family context and present from a very young age. This knowledge is informing population-level, randomized, prevention trials of family support programs.