PURPOSE: To examine the immediate and long-term effects of a walking-skill program compared with usual physiotherapy on physical function, pain and perceived self-efficacy in patients after total knee arthroplasty (TKA). METHOD: A single blind randomized controlled trial design was applied. Fifty-seven patients with primary TKA, mean age of 69 years (SD ± 9), were randomly assigned to a walking-skill program emphasizing weight-bearing exercises or usual physiotherapy. Outcomes were assessed before the interventions started at 6 weeks postoperatively (T1), directly after the interventions at 12-14 weeks (T2) and 9 months after the interventions (T3). Walking was the primary outcome, assessed by the 6 min walk test (6MWT). The secondary outcomes were timed stair climbing, timed stands, Figure-of-eight test, Index of muscle function, active knee range of motion, Knee Injury and Osteoarthritis Outcome Score and self-efficacy score. RESULTS: From T1 to T2, a better 6MWT score was found in favor of the walking-skill program of 39 m (2-76), p = 0.04. The difference between the groups in 6MWT persisted at T3, 44 m (8-80), p = 0.02. No differences in other outcome measures were found. CONCLUSION: The walking-skill program had better effect on walking than usual physiotherapy. Weight bearing was tolerated. Implications for Rehabilitation Weight-bearing exercises are tolerated by the patients in the early stage after TKA. Physiotherapy that focuses on learning different ways of walking through practice may be a plausible way to train patients after TKA.