AbstractObjective:To determine whether primary school children’s weight status and dietary behaviours vary by remoteness as defined by the Australian Modified Monash Model (MMM).Design:A cross-sectional study design was used to conduct secondary analysis of baseline data from primary school students participating in a community-based childhood obesity trial. Logistic mixed models estimated associations between remoteness, measured weight status and self-reported dietary intake.Setting:Twelve regional and rural Local Government Areas in North-East Victoria, Australia.Participants:Data were collected from 2456 grade 4 (approximately 9–10 years) and grade 6 (approximately 11–12 years) students.Results:The final sample included students living in regional centres (17·4 %), large rural towns (25·6 %), medium rural towns (15·1 %) and small rural towns (41·9 %). Weight status did not vary by remoteness. Compared to children in regional centres, those in small rural towns were more likely to meet fruit consumption guidelines (OR: 1·75, 95 % CI (1·24, 2·47)) and had higher odds of consuming fewer takeaway meals (OR: 1·37, 95 % CI (1·08, 1·74)) and unhealthy snacks (OR = 1·58, 95 % CI (1·15, 2·16)).Conclusions:Living further from regional centres was associated with some healthier self-reported dietary behaviours. This study improves understanding of how dietary behaviours may differ across remoteness levels and highlights that public health initiatives may need to take into account heterogeneity across communities.