This study aims at identifying the structural dynamic characteristics using an ambient vibration, and developing a health monitoring system which adopts damage detecting algorithms. One of the main problems for this system design is to measure long-time dynamic response signals and simultaneously estimate the structural dynamic properties. In order to be suitable for a long-time monitoring, we conduct an ambient dynamic testing for a 3-floor moment resistance steel structure and analyze structural dynamic characteristics using time domain estimation techniques. Also, a damage detecting test is performed to evaluate damage state by various detecting algorithms (modal correlation method (MAC & COMAC), eigen-parameter change method). Finally, this paper suggests the optimal algorithms for the identification of the structural damage locations and damage quantities with all such comparisons. The algorithms presented in this paper prove to be applicable in structural health monitoring of structures.