The test–retest reliability of a 1-min walk test at a child's maximum walking speed was assessed in children with bilateral spastic cerebral palsy (BSCP). Twelve male and five female children (age range 3–18 years, mean age 12 years 8 months) participated in the study. Children were classified as GMFCS level I (n=5), level II (n=8) and level III (n=4). Results showed that for walk tests performed on different days, distances varied by no more than 13.1m (for 95% of participants) and that a practice walk was vital for reducing systematic bias. The intraclass correlation coefficient was 0.94. A 1-min walk test is a reliable method of assessing function in children with CP but care must be taken when interpreting changes in individual patient data.