Damage to foundation structures which cause inclination/settlement of buildings resulting in prohibition of the usage has been reported after recent earthquakes in Japan. Such damage was often observed to boundaries between superstructure and foundation structure with piles. In order to make judgement of availability of earthquake-damaged buildings with pile foundation, assessment of piles is indispensable; however, there are very few experimental researches clarifying the behavior of column-foundation beam-pile joints which are subjected to complicated seismic loads. Therefore, a series of experimental tests for the column, foundation beam and cast-in-place pile joints under realistic seismic loads was conducted to obtain the fundamental data, which is likely to provide key behavior to detect damage to the pile foundations after seismic events. In this study, a SRC building damaged to the foundation beams, piles, and pile caps in the 1995 Kobe earthquake was focused (Figs. 1 to 3). In particular, an exterior column-foundation beam-pile joint with severe damage was targeted. Shear forces applied to the exterior pile are affected by complicated seismic actions, namely, inertial forces at the foundation slab and beam as well as shear and variable axial forces to the exterior column. Therefore, numerical analyses using several models considering superstructure-pile interactions (Fig. 4, Tables1-4) were conducted to evaluate the seismic loads applied to the target joint. As a result, the inertial force at foundation was approximately 0.5 times the column shear force (Figs. 5 to 8). Based on the analytical findings, an experimental method for exterior column-foundation beam-pile joints simulating the realistic seismic behavior was proposed and applied to the structural tests. Four 1/3-scale specimens were prepared considering two parameters which were the existence of a superstructure wall on the foundation beam and the volume of the pilecap (Figs. 9-10, Tables5-7). The specimens were supported with a pin at the inflection points of the column and pile, respectively, and a roller at that of the beam. Horizontal cyclic loads were applied to the ends of column and foundation beam with a ratio of 1:0.5 simulating realistic shear distributions to the column and pile (Figs. 11-12). Variable axial loading was also applied to the column (pile) proportional to the applied shear force. In the cases of the specimens without the wall, the stiffness significantly decreased with flexural yielding at the end of foundation beam and the top of pile during the loading cycle to 0.5 ×10-2rad in the positive and negative directions, respectively (Fig. 13). On the other hands, in the cases of the specimens with the wall, the strengths rapidly dropped by concrete crushing at the top of pile during the loading cycle to 1.5 ×10-2rad. In particular, shear cracks on the pilecap were observed in the specimen with smaller pilecap representing the target building; however, the damage was less severe than that observed in the building. Typical bending analyses considering the post-peak strength deterioration for concrete well simulated the test results (Figs. 14-15). Furthermore, strain decreases of the beam longitudinal rebar were observed with the concrete of pile crushing (Fig. 16), which may provide a scheme to detect damage to pile foundations. Consequently, the fundamental behavior/performance of the exterior column-foundation beam-pile joints were experimental evaluated, which contributes to assess damage to piles as well as to improve the seismic design.