Background: The motorized treadmill elicits stepping in infants, however the effects of belt speed on treadmill-elicited stepping patterns are still unknown. Objective: To examine the effect of belt speed on treadmill-elicited infant stepping. Methods: Six normally developing infants, aged between 11 and 13 months and acquiring independent walking, were videotaped while stepping on a treadmill at four belt speeds (0.1, 0.16, 0.22, and 0.28 m/s), with passive markers affixed to the major joints (shoulder, hip, knee, and ankle). The images were digitized and analyzed using the software programs Ariel Performance Analysis System, Matlab, and SPSS (Statistical Package for Social Sciences, v.10). The analyzed variables were the descriptive, temporal, coordinative, and angular characteristics of the treadmill-elicited steps at the four speeds. Results: At 0.22 and 0.28 m/s, stride velocity and duration increased, and the duration of the stance phase decreased at 0.22 m/s. Moreover, at 0.22 and 0.28 m/s, the knee joint range of motion increased. Conclusions: The moderate and high speeds seem to be more appropriate to elicit stepping in infants who are acquiring independent walking, with 0.22 m/s being the speed that seemed the most appropriate. Article registered in the Australian New Zealand Clinical Trials Registry (ANZCTR) under number ACTRN12609000517279.