This paper develops a system identification process based on seismic response data to investigate the physical parameters of an isolated highway bridge equipped with the lead-rubber bearings (LRBs). The effects of soil-structure interaction (SSI) of the highway bridge built on pile foundations are taken into account in the analysis. Unlike many other identification methods that generally seek for the equivalent system parameters or modal parameters of the target structure, the proposed method aims at identifying physical parameters such as the pier stiffness or hysteresis of the LRB that are considered more critical and useful for health monitoring of bridges. In this study, a linear elastic model is used for the pier while a bilinear model is considered to characterize the hysteretic behavior of the LRB. The bilinear hysteretic model is in turn reduced to a backbone curve based on the Masing criterion by which the multi-valued restoring force is transformed into a single-valued function so that the computational effort of the identification process can be minimized. Numerical examples have been conducted, including the identification of Bai-Ho Bridge, which is the first seismic isolation bridge in Taiwan, to demonstrate the feasibility of the proposed technique for dynamic systems with inelastic components and SSI.