Aim: To determine if robotic assisted gait training (RAGT) using surface muscle electrical stimulation and locomotor training enhances mobility outcomes when compared to locomotor training alone in children with cerebral palsy (CP). Method: Forty children (18 females, 22 males; mean age 8y 1mo, SD 2y 1mo; range 5y 1mo–12y 11mo) with CP in Gross Motor Function Classification System levels (GMFCS) III, IV, and V were randomly assigned to the RAGT and locomotor training (RAGT+LT) group or locomotor training only group (dosage for both: three 1-hour sessions a week for 6 weeks). Outcomes were assessed at baseline T1 (week 0), post-treatment T2 (week 6), and retention T3 (week 26). The primary outcome measure was the Goal Attainment Scale. Secondary outcome measures included the 10-metre walk test, children's functional independence measure mobility and self-care domain, the Canadian Occupational Performance Measure, and the Gross Motor Function Measure. Results: There were no significant differences between the groups for both the primary and secondary outcome measures. All participants completed the intervention in their original group allocation. There were no reported adverse events. Interpretation: The addition of RAGT to locomotor training does not significantly improve motor outcomes in children with CP in GMFCS levels III, IV, and V. Future studies could investigate health and well-being outcomes after locomotor training. What this paper adds: Marginally ambulant and non-ambulant children with cerebral palsy can participate in locomotor training. Robotic assisted gait training when added to locomotor training does not appear to be any more effective than locomotor training alone.