||Four main factors may contribute to the spatial variation of earthquake excitations, i.e. coherency effect, attenuation effect, site effect and propagation effect. Seismic response of long span bridges including the effects of spatial variation of seismic waves is studied in this thesis. The research work contains two parts: shake table model test and numerical analysis. A dual shake table system is developed for the purpose of this thesis research. The general concept of scaling law (similitude requirements) is stated. To estimate the required capacity of the proposed dual shake table, reduced scale models of different types of bridges are fictitiously designed based on the similitude requirements. After the construction, the verification test of the dual shake table system is carried out. A reduced scale model of a long span cable-stayed bridge is designed for shake table tests. The prototype is the Kap Shui Mun Bridge in Hong Kong. The derivation of scale factor table for model design and the model installation procedure are introduced. The earthquake records from the SMART-1 and the PEER Strong Motion Database are used in this test. The numerical processing scheme suggested by the USGS is adopted to process the raw records. The test cases include modal test, identical-input excitation test and varying-input excitation test. The numerical analysis of the long span cable-stayed bridge subjected to non-uniform excitation is performed. A refined 3-D finite element model of the Kap Shui Mun Bridge is developed. The nonlinear time-history analyses of the FEM model subjected to selected non-uniform excitation cases are performed. The numerical analysis results are compared with the ones from shake table testing, in order to verify the numerical method as well as shed some light on the effects of spatially varying earthquake ground motion on the seismic behavior of long span bridges.