Introduction No prospective studies have examined the prevalence, antecedents or concurrent characteristics associated with self-harm in non-treatment-seeking primary school-aged children. Methods In this cohort study from Melbourne, Australia we assessed 1239 children annually from age 8–9 years (wave 1) to 11–12 years (wave 4) on a range of health, social, educational and family measures. Past-year self-harm was assessed at wave 4. We estimated the prevalence of self-harm and used multivariable logistic regression to examine associations with concurrent and antecedent factors. Results 28 participants (3% of the 1059 with self-harm data; 18 girls [3%], 10 boys [2%]) reported self-harm at age 11–12 years. Antecedent (waves 1–3) predictors of self-harm were: persistent symptoms of depression (sex-age-socioeconomic status adjusted odds ratio [aOR]: 7.8; 95% confidence intervals [CI] 2.6 to 24) or anxiety (aOR: 5.1; 95%CI 2.1 to 12), frequent bullying victimisation (aOR: 24.6; 95%CI 3.8 to 158), and recent alcohol consumption (aOR: 2.9; 95%CI 1.2 to 7.1). Concurrent (wave 4) associations with self-harm were: having few friends (aOR: 8.7; 95%CI 3.2 to 24), poor emotional control (aOR: 4.2; 95%CI 1.9 to 9.6), antisocial behaviour (theft—aOR: 3.1; 95%CI 1.2 to 7.9; carrying a weapon—aOR: 6.9; 95%CI 3.1 to 15), and being in mid-puberty (aOR: 6.5; 95%CI 1.5 to 28) or late/post-puberty (aOR: 14.4; 95%CI 2.9 to 70). Conclusions The focus of intervention efforts aimed at preventing and reducing adolescent self-harm should extend to primary school-aged children, with a focus on mental health and peer relationships during the pubertal transition.