Ankle–foot orthoses (AFOs) are commonly prescribed to children with cerebral palsy (CP). The conventional AFO successfully controls the first and second ankle rocker, but it fails to correct the third ankle rocker, which negatively effects push-off power. The current study evaluated a new powered AFO (PAFO) design, developed to address the shortcomings of the conventional AFO. Eight children with spastic CP (12.4 ± 3.4 years; GMFCS I-III; 4/4-♂/♀; 3/5-bi/unilateral) were included. Sagittal kinematic and kinetic data were collected from 20 steps during barefoot walking, with conventional AFOs and PAFOs. In the PAFO-condition, an actuation unit was attached to a hinged AFO and through push–pull cables to a backpack that was carried by the child and provided patient-specific assistance-as-needed. SnPM-analysis indicated gait cycle sections that differed significantly between conditions. For the total group, differences between the three conditions were found in ankle kinematics (49.6–66.1%, p = 0.006; 88.0–100%, p = 0.011) and angular velocity (0.0–6.0%, p = 0.001; 45.1–51.1%, p = 0.006; 62.2–73.0%, p = 0.001; 81.2–93.0%, p = 0.001). Individual SnPM-analysis revealed a greater number of significant gait cycle sections for kinematics and kinetics of the ankle, knee, and hip. These individual results were heterogeneous and specific per gait pattern. In conclusion, the new PAFO improved the ankle range-of-motion, angular velocity, and power during push-off in comparison to the conventional AFO.