Falls are among the leading causes of fatal and non-fatal injuries for seniors. For long-term non-contact detection of falls in home and nursing environments, a coherent frequency-modulated continuous-wave (FMCW) radar sensor is designed and tested. The coherence property of the developed FMCW radar helps to preserve the phase history of the signal detected from human subjects, and can thus obtain Doppler information based on inverse synthetic aperture radar (ISAR) imaging. By analyzing the changes of the radar cross-section (RCS), range, and Doppler on ISAR images during the movement of a subject, falls can be distinguished from normal activities. Experiments have been carried out to demonstrate and analyze the ISAR signature of actual fall incidents, which has been compared with that of an abrupt movement. Preliminary results have confirmed the ability of an FMCW radar to successfully detect fall events using ISAR imaging.