The aim of this study was to determine whether breastfeeding in early life is associated with bone mass in prepubertal children. We studied 330 8-year-old male and female children from Southern Tasmania representing 47% of those who originally took part in a birth cohort study of risk factors for Sudden Infant Death Syndrome in 1988. Breastfeeding intention and habit were assessed in both 1988 and 1996. Bone mineral density was measured by dual-energy X-ray densitometry. Children who were breastfed had higher bone mineral density at the femoral neck (+0.20 SD, p = 0.07), lumbar spine (+0.25 SD, p = 0.03) and total body (+0.29 SD, p = 0.006) compared with those who were bottle-fed. This association with breastfeeding was present in children born at term (femoral neck: +0.26 SD, p = 0.05; lumbar spine: +0.34 SD, p = 0.007; total body: +0.41 SD, p = 0.0008) but not those born preterm, and remained significant after adjustment for size, lifestyle factors and socioeconomic factors. Breastfeeding for less than 3 months was not associated with increased bone mass at any site. In conclusion, this study has demonstrated a beneficial association between breastfeeding in early life and bone mass in 8-year-old children born at term, particularly those breastfed for 3 months or longer, which appears biological. If this association is confirmed in other populations and persists until the attainment of peak bone mass then the implication would be that osteoporosis prevention programs need to start very early in the life cycle.