Objective:To compare the effectiveness of two types of robotic training in improving the lower extremity motor functioning of children with spastic cerebral palsy (SCP).Methods:Twenty-eight children with SCP were randomly divided into a control group and an experimental group, each of 14. Both groups received conventional exercise therapy, paraffin therapy, neuromuscular electrical stimulation, and massage. Both also performed 30 minutes of gait training five days a week for eight weeks assisted by either a Lokomat or a Relink lower limb rehabilitation robot. Before and after the treatment, both groups were evaluated using the Gross Motor Function Measure (GMFM), the Pediatric Balance Scale (PBS), the Modified Tardieu Scale (MTS), the six-minute walk test (6MWT), the Physiological Cost Index (PCI) and their self-selected walking speed (SWS).Results:Significant improvement in all of the measurements were observed in both groups. After the treatment, there were no significant differences between the two group in the average GMFM (section D and E) or PBS scores. The average MTS R1 and R2, SWS, 6MWT and PCI results of the experimental group were, however, significantly better than those of the control group.Conclusion:Applying either the Lokomat or Relink robot in lower extremity rehabilitation improves the lower extremity motor function of children with grade II-III SCP. The Relink robot is the more effective in improving triceps surae spasm and walking ability.