This research investigates a rapid structural health monitoring method for civil structures that delivers results immediately after a major seismic event. The method is based on an overall least squares regression analysis and hypothesis test, and is applied to an experimental scaled 12-story reinforced concrete frame structure subjected to two stages of input ground motions. The test structure is simplified to a six degree of freedom (DOF) system model, where the accelerations of each DOF are recorded. Nonlinear hysteresis loops are reconstructed for each DOF, and divided into a number of half cycles in chronological order. Changes in identified elastic story stiffness in these selected half cycles are used as an index for damage localization and severity assessment, and tracked over time using the proposed identification algorithm. The feasibility and robustness of the proposed monitoring algorithm is validated using the data from the experimental test structure. No large drop of elastic stiffness was identified for the test structure under the small input ground motion. Significant stiffness degradation was identified for the fourth, third and second DOF with stiffness losses over 50% compared to the calculated initial stiffness. Cracks were also observed at the beam-column joint connection at floors six, five and four after the experiment for the strong input ground motion. The results indicate the identification algorithm is capable of detection and assessing the damage location and severity automatically by tracking the evolution of elastic story stiffness without requiring human input.