<p>Robot-assisted gait training may improve walking recovery after total knee arthroplasty (TKA). However, optimal intervention durations and early postoperative feasibility remain unclear. We evaluated a short-term, early postoperative gait-training program’s effects on walking-speed recovery using a lightweight, hip-wearable assistive device. In this pilot study, patients who underwent TKA were allocated preoperatively to receive either HWA-01 gait training plus standard rehabilitation (HGT) or standard rehabilitation alone (control). HGT was delivered during postoperative weeks 1–3 (8–10 sessions; ≤ 20&#xa0;min/session, excluding rest). Outcomes were assessed at baseline, 2, 4, and 8&#xa0;weeks; generalized linear mixed models were used to analyze gait speed, knee function, and Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) outcomes, and Patient Acceptable Symptom State (PASS) achievement for self-selected walking speed (SWS) (≥ 1.2&#xa0;m/s) was compared between groups. Forty patients (52 knees) were analyzed (HGT, 13 patients/13 knees; control, 29 patients/39 knees). Significant time × group interactions were observed at 2&#xa0;weeks for SWS (<i>p</i> = 0.045) and maximum walking speed (<i>p</i> = 0.044), indicating an attenuation of the early postoperative walking-speed decline in the HGT group. At 4&#xa0;weeks, more patients in the HGT group achieved the PASS threshold (<i>p</i> = 0.013). No significant group or interaction effects were observed for the range of motion, torque, or WOMAC outcomes. A 2-week, early postoperative HGT program using a lightweight hip-assist device may mitigate early walking-speed decline after TKA and promote earlier attainment of a patient-acceptable walking speed with favorable safety and tolerability.</p><p><?qj left?><?noindent??><i>Trial registration</i>: UMIN Clinical Trials Registry (UMIN-CTR), UMIN000054889, <a href="https://www.umin.ac.jp/ctr/">https://www.umin.ac.jp/ctr/</a>.</p>