Buildings crossing active faults often suffer severe damage due to fault dislocation during direct-type urban earthquakes. This study employs physical model tests to systematically investigate the dynamic response mechanisms of the integrated “surface rupture zone–overburden–foundation–superstructure” system subjected to bedrock dislocation. A testing apparatus capable of simulating reverse faults with adjustable dip angles (45° and 70°) was developed. Using both sand and clay as representative overburden materials, the experiments simulated the processes of surface rupture evolution, foundation deformation, and structural response under varying fault dislocation magnitudes. Results indicate that the fault rupture pattern is governed by the bedrock dislocation magnitude, soil type, and fault dip angle. The failure process can be categorized into three distinct stages: initial rupture, rupture propagation, and rupture penetration. The severity and progression of structural damage are primarily determined by the building’s location relative to the fault trace. Structures located entirely on the hanging wall exhibited tilting angles that remained below the specified code limit throughout the dislocation process, demonstrating behavior dominated by rigid-body translation. In contrast, buildings crossing the fault exceeded this limit even at low dislocation levels, developing significant tilt and strain concentration due to differential foundation settlement. The most severe damage occurred in high-angle dip sand sites, where the maximum structural tilt reached 5.5°. This research elucidates the phased evolution of seismic damage in straddle-fault structures, providing experimental evidence and theoretical support for the seismic design of buildings in near-fault regions. The principal theoretical and methodological contributions are (1) developing a systematic “fault–soil–structure” testing methodology that reveals the propagation of fault dislocation through the system; (2) clarifying the distinct failure mechanisms between straddle-fault and hanging-wall structures, providing a quantitative basis for targeted seismic design; and (3) quantifying the controlling influence of fault dip angle and soil type combinations on structural damage severity, identifying high-angle dip sand sites as the most critical scenario.

To investigate the performance of horizontally reinforced composite foundations in resisting surface rupture of normal faults, this study designed and conducted a series of physical model tests. A systematic comparative analysis was performed on the fracture resistance of sites with three-layer sand, five-layer sand, and three-layer clay geogrid horizontally reinforced composite foundations under 70° normal fault dislocation. The results indicate that significant changes in earth pressure serve as a precursor indicator of fault rupture, and their evolution process reveals the internal energy accumulation and release mechanism. Increasing the number of geogrid layers significantly enhances the lateral confinement of the foundation, resulting in a narrower macro-rupture zone located farther from the structure in sand sites, and promotes the formation of a step-fault scarp deformation mode at the surface, which is more conducive to structural safety. Under identical reinforcement conditions, the clay site exhibited comprehensively superior fracture resistance compared to the sand site due to the soil cohesion and stronger interfacial interaction with the geogrids, manifested as more significant deviation of the rupture path, and lower microseismic accelerations and structural strains transmitted to the building. Comprehensive analysis confirms that employing geogrid-reinforced composite foundations can effectively guide the surface rupture path and improve the deformation pattern, representing an effective engineering measure for mitigating disaster risk for buildings spanning active faults.