Measurement and modeling of low frequency ground reaction forces (GRFs) from human walking have been the subjects of research in different fields from biomechanics to civil engineering and structural dynamics. Many of the existing models are developed based on experiments which alter natural walking by for example presence of force transducers, limitations in the speed and path of walking and replacing the real floor with a transducerfacilitated measurement rig. These alterations result in contact forces which do not represent real GRFs. In this study, a time-domain inverse measurement method based on LMS algorithm is used to measure low-frequency (<100 Hz) forces induced by human footsteps. The LMS-based force identification method is first validated for low-frequency excitations with less complexity in number of excitation positions and frequency content compared with the footsteps. The method is then applied to measure ground reaction forces created by human walking.