Several studies indicated that walking with an ankle foot orthosis (AFO) impaired third rocker. The purpose of this study was to evaluate the effects of two types of orthoses, with similar goal settings, on gait, in a homogeneous group of children, using both barefoot and shoe walking as control conditions. Fifteen children with hemiplegia, aged between 4 and 10 years, received two types of individually tuned AFOs: common posterior leaf-spring (PLS) and Dual Carbon Fiber Spring AFO (CFO) (with carbon fibre at the dorsal part of the orthosis). Both orthoses were expected to prevent plantar flexion, thus improving first rocker, allowing dorsiflexion to improve second rocker, absorbing energy during second rocker, and returning it during the third rocker. The effect of the AFOs was studied using objective gait analysis, including 3D kinematics, and kinetics in four conditions: barefoot, shoes without AFO, and PLS and CFO combined with shoes. Several gait parameters significantly changed in shoe walking compared to barefoot walking (cadence, ankle ROM and velocity, knee shock absorption, and knee angle in swing). The CFO produced a significantly larger ankle ROM and ankle velocity during push-off, and an increased plantar flexion moment and power generation at pre-swing compared to the PLS (<0.01). The results of this study further support the findings of previous studies indicating that orthoses improve specific gait parameters compared to barefoot walking (velocity, step length, first and second ankle rocker, sagittal knee and hip ROM). However, compared to shoes, not all improvements were statistically significant.