Damage process monitoring of concrete structures using acoustic emission (AE) technology has been drawn more and more attention due to its powerful capability. In this paper, a brand new cement-based piezoelectric composite sensor was introduced with improved performance, particularly concerning its AE signals detection capacity in the high frequency domain. Such sensors were embedded into the foundation of reinforced concrete (RC) frames during construction, and appointed to monitor concrete damage due to shake table excitation. A standard ground motion record of 1999 Taiwan earthquake with 840 gal and 1300 gal peak ground accelerations were employed in the tests as the excitation inputs. The signals detected by the sensors were stored and analyzed by commercial available data acquisition devices. Advanced stochastic signal analysis methods were adopted to effectively interpret the frequency domain components and identify the useful information representing the damage processes of the RC frames. The results were compared with the eye observations of structural damage and corresponding cyclic loading tests results. It is shown that the health monitoring method using cement-based piezoelectric composite sensors and advanced stochastic signal analysis are capable of detecting and evaluating the damage process of RC frames due to seismic loading. An effective damage indicator of the RC frames is possible to be evaluated from extracted AE information. © RILEM 2010.