To investigate the vibration isolation effect of composite vibration isolation walls on surface vibrations in suburban railway deep tunnels under various influencing factors, an integrated numerical model of the train was initially developed. This model solved the wheel-rail interaction force and was applied to a three-dimensional volume coupling model of the track soil. Subsequently, the model's reliability was validated through comparison with measured data. Afterward, the vibration isolation effects of various types of EPS material vibration isolation walls were examined, with a focus on exploring the impact of thickness, material proportion, and relative positioning of the materials within the vibration isolation wall composed of EPS material and concrete. Research indicates that with an increase in the burial depth of a single material vibration isolation wall, its effective vibration isolation frequency range gradually widens. When the burial depth of the vibration isolation wall exceeds the tunnel burial depth, the vibration isolation effect is optimal. Composite vibration isolation walls, with thicknesses smaller than single-material vibration isolation walls, exhibit superior vibration isolation effects compared to their single-material counterparts. The effective vibration isolation frequency band of composite vibration isolation walls differs from that of single-material vibration isolation walls. Using the optimal-size vibration isolation wall of a single material as a composite vibration isolation wall enhances the vibration isolation effect of peak acceleration in the frequency domain by 16.58% and peak velocity by 16.95%. Moreover, frequency domain peak displacement experiences a 30.73% improvement in the vibration isolation effect.