BACKGROUND AND OBJECTIVES: In people with hereditary spastic paraplegia (HSP), reduced gait adaptability is common and disabling. Gait impairments result from lower extremity spasticity, muscle weakness, and impaired proprioception. The aim of this study was to assess the efficacy of a 5-week gait-adaptability training in people with pure HSP. METHOD: We conducted a randomized clinical trial with a cross-over design for the control group, and a 15-week follow-up period after training. Thirty-six people with pure HSP were randomized to 5 weeks of (i) gait-adaptability training (10 hours of C-Mill training-a treadmill equipped with augmented reality) or (ii) a waiting-list control period followed by 5 weeks gait-adaptability training. Both groups continued to receive usual care. The primary outcome was the obstacle subtask of the Emory Functional Ambulation Profile. Secondary outcome measures consisted of clinical balance and gait assessments, fall rates, and spatiotemporal gait parameters assessed via 3D motion analysis. RESULTS: The gait-adaptability training group (n = 18) did not significantly decrease the time required to perform the obstacle subtask compared to the waiting-list control group (n = 18) after adjusting for baseline differences (mean: -0.33 seconds, 95% CI: -1.3, 0.6). Similar, non-significant results were found for most secondary outcomes. After merging both groups (n = 36), the required time to perform the obstacle subtask significantly decreased by 1.3 seconds (95% CI: -2.1, -0.4) directly following 5 weeks of gait-adaptability training, and this effect was retained at the 15-week follow-up. CONCLUSIONS: We found insufficient evidence to conclude that 5 weeks of gait-adaptability training leads to greater improvement of gait adaptability in people with pure HSP.