Aim: To investigate the relationship between motor ability and early social development in a cohort of preschool children with cerebral palsy (CP). Design: Population-based cohort study. Methods: Participants were 122 children with CP assessed at 18, 24 and 30 months, corrected age (ca). Motor ability was measured by the Gross Motor Function Classification System (GMFCS) with classification assigned by physiotherapists. The sample was representative of a population-based cohort (I=48, 38.4%, II=19, 15.2%, III=17, 13.6%, IV=22, 17.6% and V=19, 15.2%). Social development was measured by the Paediatric Evaluation of Disability Inventory (PEDI) and included capabilities in social interaction, social communication, interactive play and household/community tasks. Results: Cross-sectional analyses indicated a significant relationship between motor ability and social development at 18 months, F(4, 56) = 11.44, p less than 0.0001, [eta][superscript 2] = 0.45, at 24 months, F(4, 79) = 15.66, p less than 0.0001, [eta][superscript 2] = 0.44 and at 30 months, F(4, 76) = 16.06, p less than 0.0001, [eta][superscript 2] = 0.49. A longitudinal analysis with a subset of children (N = 24) indicated a significant interaction between age at assessment and GMFCS, F(2, 21) = 7.02, p = 0.005, [eta][superscript 2] = 0.40. Comparison with community norms indicated that at 18 months corrected age, 44.3% of the cohort was greater than two standard deviations below the mean (greater than 2SD) for social development and a further 27.9% of the cohort was greater than one standard deviation below the mean (greater than 1SD). Interpretation: There is a relationship between motor ability and social development in preschool children with CP. Children with CP may require support for social development in additional to physical interventions, from as early as 18 months. (Contains 3 tables and 1 figure.)