BACKGROUND: Identifying context-specific correlates of home- and neighbourhood-based physical activity in preschool-aged children may help improve intervention program development for these settings. METHODS: A total of 153 3-4-year-old children were recruited through preschool settings in Cambridgeshire (January-July 2013). Children wore Actiheart accelerometers for ≤7 days to assess their sedentary time (ST), light-(LPA) and moderate- to vigorous-intensity physical activity (MVPA). A parent-completed questionnaire assessed correlates across the ecological model and the child's preschool attendance during the measurement week. Only accelerometer data for times when children were at home were used. Multilevel models (Level 1: days; Level 2: child) examined associations between maternal-reported exposure variables and each outcome (children's home- and neighbourhood-based ST, LPA and MVPA) (main analysis). Further analyses included the subsample of children with complete paternal correlates data (father analysis). RESULTS: In the main analyses, children with older siblings engaged in less ST. Children whose mothers reported being 'moderately inactive' or 'active' (vs. inactive) engaged in less LPA, while children whose mothers worked >35 h week-1 engaged in less MVPA. More equipment at home was associated with lower LPA but greater MVPA. In the father analysis, father's television viewing before 6 pm was associated with greater ST and less MVPA in children; the negative association between mother's activity and children's LPA was retained. CONCLUSION: Family demographics and parental behaviours appear to have the strongest association with children's home- and neighbourhood-based ST, LPA and MVPA. This study further highlights the importance of examining both maternal and paternal behaviours.