The aim of this study was to determine the reliability of family and professional assessment of manual ability using the Manual Ability Classification System (MACS) for children with cerebral palsy (CP) in the UK. Families who were taking part in a study measuring the activities and participation of children with CP were invited to classify their child's manual ability using the MACS. Postal surveys were conducted with the families and health professionals nominated by the families. Perfect agreement was assessed as a percentage; chance-corrected agreement was measured using Cohen's kappa (κ), and reliability was determined using the intraclass correlation coefficient (ICC). Families of 91/128 (71%) children responded to the survey (53 males, 38 females; mean age 9y 11mo [SD 1y 11mo], range 6–12y) out of whom 88 indicated a single MACS level. Seventy-two children (82%) were classified with spastic CP, 12 (14%) with dyskinesia, two (2%) with ataxia, and two (2%) were not classified. There were 21, 27, 11, 10, and 19 children who were classified by their families in Gross Motor Function Classification System Levels I to V respectively; 14, 30, 18, 13, and 13 children classified by their families in MACS levels I to V. The survey of health professionals generated 60/71 (85%) responses from physiotherapists, 55/58 (93%) responses from paediatricians, and 21/24 (88%) responses from occupational therapists. There was perfect agreement between families and professionals for more than 50% of children; the indices of chance-corrected agreement ranged from κ=0.3 to 0.5, and the reliability coefficients ranged from ICC 0.7 to 0.9. Indices of agreement and reliability between families and professionals were equivalent to those between different professionals. The MACS, therefore, offers a valid and reliable method for communicating about the manual ability of children with CP. Families and professionals may not always agree precisely on a MACS level, particularly if children's performance of manual tasks varies in different environments.