New surface electromyogram (SEMG) techniques offer the potential to advance knowledge of healthy and diseased motor units. Conduction velocity (CV) estimates, obtained from indwelling electrodes, may provide diagnostic information, but the standard method of CV estimation from SEMG may be of only limited value. We developed a motor unit (MU) tracking algorithm to extract motor unit conduction velocity (MUCV) and motor unit action potential (MUAP) amplitude estimates from SEMG. The technique is designed to provide a noninvasive means of accessing fatigue and recruitment behavior of individual MUs. We have applied this MU tracking algorithm to SEMG data recorded during isometric fatiguing contractions of the tibialis anterior (TA) muscle in nine healthy subjects, at 30%-40% maximum voluntary contraction (MVC). The results reveal that MUCVs and MUAP amplitudes of individual MUs can be estimated and tracked across time. Time-related changes in the MU population may also be monitored. Thus, the SEMG technique employed provides insight into the behavior of the underlying muscle at the MU level by noninvasive means. © 2005 Wiley Periodicals, Inc.