Background: Gait disruption is a common poststroke problem. Robot-assisted gait training (RAGT) might improve motor function, balance, and activities of daily living. Objective: We compared the clinical effectiveness of early integrated RAGT using the Walkbot robotic gym with an intensity-matched enhanced lower limb therapy (ELLT) program and with conventional rehabilitation therapy (CRT) in patients with acute ischemic stroke. Methods: A total of 192 patients with acute ischemic stroke were randomly assigned (1:1:1) to receive RAGT, ELLT, or CRT. All three groups received 45 min of training daily, 3 days a week, for 4 weeks consecutively. Before and after the 4-week treatment, the patients were assessed based on a 6-minute walking test (6MWT), functional ambulation classification (FAC), timed up and go (TUG) test, dual-task walking (DTW) test, Tinetti's test, Barthel's index (BI), stroke-specific quality of life (SS-QOL) scale, and gait analysis parameters. Results: After the 4-week intervention, the results of the 6MWT, FAC, TUG, DTW, Tinetti's test, BI, SS-QOL, and gait in the three groups significantly improved. Compared with ELLT and CRT groups, participants in the RAGT group had a better performance in 6MWT (199.11 +/- 60.72 versus 182.47 +/- 59.72 versus 173.69 +/- 40.58, p = 0.035), FAC (4.10 +/- 0.91 versus 3.69 +/- 0.88 versus 3.58 +/- 0.81, p = 0.044), DTW (10.29 +/- 2.38 versus 12.92 +/- 2.64 versus 13.89 +/- 2.62, p = 0.031), SS-QOL (184.46 +/- 20.53 versus 165.39 +/- 20.49 versus 150.72 +/- 20.59, p = 0.012), velocity (0.66 +/- 0.22 versus 0.55 +/- 0.23 versus 0.51 +/- 0.20, p = 0.008), cycle duration (1.38 +/- 0.40 versus 1.50 +/- 0.38 versus 1.61 +/- 0.30, p = 0.040), and swing phase symmetry ratio (SPSR, 1.10 +/- 0.33 versus 1.21 +/- 0.22 versus 1.48 +/- 0.25, p = 0.021). The TUG, Tinetti's test, BI, and RMT results were similar, however. Conclusion: In the acute stroke phase, early integrated RAGT showed greater performance in gait rehabilitation than CRT and ELLT. Registration: ChiCTR1900026225