There is limited scientific knowledge on ACL injuries in children 12 years or younger. Substantial controversy exists on treatment algorithms and there are no published data on performance-based functional outcome. Classification of adult ACL injured subjects as copers and non-copers is common, but no study has classified knee function in children using performance-based functional test after ACL injury. The aim of the present study was to evaluate the medium-term functional outcome among children with ACL injury and to classify them as copers and non-copers. Children 12 years or younger who were referred to our institution from 1996 to 2004 with an ACL injury were included. Twenty non-operated subjects (21 knees) and six ACL reconstructed subjects (7 knees) were examined at a minimum of 2 years after ACL injury or reconstruction. Four single-legged hop tests, isokinetic muscle strength measurements, and three functional questionnaires (IKDC 2000, KOS-ADLS and Lysholm) were used as outcome measurements. Children who had resumed their pre-injury activity level and performed above 90% on all hop tests were classified as copers following non-operative treatment and ACL reconstruction. The 26 children were on average 10.1 years at the time of injury. Of the non-operated children, 65% had returned to pre-injury activity level, and 50% were classified as copers. Copers scored significantly better than non-copers on single hop for distance, IKDC 2000, and Lysholm score. Of the non-operated children, 9.5% had suffered a secondary meniscus injury. Of the ACL reconstructed subjects, 67% were classified as copers at follow-up. Non-operated ACL-deficient children demonstrated excellent knee function on performance-based single-legged hop tests and 65% had returned to pre-injury activity level. Delayed ACL reconstruction resulted in success for a majority of the ACL-reconstructed children. Treatment algorithms for ACL-injured children are discussed.