BACKGROUND AND OBJECTIVE: Balance and walking capacity are often impaired in people with motor incomplete spinal cord injury (iSCI), frequently resulting in reduced functional ambulation and participation. This study aimed to assess the efficacy of walking adaptability training compared to similarly dosed conventional locomotor and strength training for improving walking capacity, functional ambulation, balance confidence, and participation in ambulatory people with iSCI. METHODS: We conducted a 2-center, parallel-group, pragmatic randomized controlled trial. Forty-one people with iSCI were randomized to 6 weeks of (i) walking adaptability training (11 hours of Gait Real-time Analysis Interactive Lab (GRAIL) training-a treadmill in a virtual reality environment) or (ii) conventional locomotor and strength training (11 hours of treadmill training and lower-body strength exercises). The primary measure of walking capacity was maximal walking speed, measured with an overground 2-minute walk test. Secondary outcome measures included the Spinal Cord Injury Functional Ambulation Profile (SCI-FAP), the Activities-specific Balance Confidence (ABC) scale, and the Utrecht Scale for Evaluation of Rehabilitation-Participation (USER-P). RESULTS: No significant difference in maximal walking speed between the walking adaptability (n = 17) and conventional locomotor and strength (n = 18) training groups was found 6 weeks after training at follow-up (-0.05 m/s; 95% CI = -0.12-0.03). In addition, no significant group differences in secondary outcomes were found. However, independent of intervention, significant improvements over time were found for maximal walking speed, SCI-FAP, ABC, and USER-P restrictions scores. TRIAL REGISTRATION: Dutch Trial Register; Effect of GRAIL training in iSCI.