The aim of this study was to determine the extent of ankle muscle weakness in children with cerebral palsy (CP) and to identify potential causes. Maximal voluntary contractions of plantar (PF) and dorsiflexors (DF) were determined at optimal angles in knee flexion and extension in both legs of 14 children with hemiplegia (7 males, 7 females) and 14 with diplegia (8 males, 6 females). Their results were compared to 14 age- and weight-matched control participants (5 males, 9 females). Muscle cross-sectional areas of soleus, posterior, and anterior compartment muscles were determined from MRIs in 14 children with CP (eight diplegia, six hemiplegia) and 18 control children. Specific tension (torque/unit area) of PF and DF was determined from torque and cross-sectional area results. Muscle volumes of PF and DF were also determined in both legs of five control children and five with hemiplegia. Muscle EMG was recorded from soleus, medial gastrocnemius, and tibialis anterior during each maximal voluntary contraction. Mean amplitude was significantly reduced in PF and DF in both CP groups and significantly higher levels of coactivation of antagonists were found compared to control participants. Strength of PF and DF was significantly reduced in both CP groups, but more importantly the muscles were found to be weak based on significantly reduced specific tensions. The PF were most affected, particularly in the group with hemiplegia. It is believed that an inability to maximally activate their muscles contributed to this weakness. A combination of incomplete activation and high levels of PF coactivation are thought to have contributed to DF weakness.