Wheel polygon wear (WPW) is a common form of wheel non-circularity in high-speed trains. As the train speed increases, the WPW and track irregularity excitation become coupled, which not only excites the structural deformation of the flexible wheelset, but also aggravates the high-frequency components of the axle box’s vibration response. This paper develops a rigid-flexible coupling dynamic model using the Craig-Chang reduction method, and analyzes the coupling effect of track irregularity excitation and WPW on the axle box vibration response at 300 km/h. The study also investigates the influence of WPW wear order and depth on the amplitude of the response harmonic component, the amplitude of the wheelset modal frequency, and the safety evaluation metrics under the coupling effects. The results show that the coupling effect significantly increases the vibration level of the axle box, but reduces the amplitude of the wheel bending, torsional deformation modal frequency and the WPW excitation frequency. This study provides theoretical and technical support for the dynamic optimization of high-speed train wheel groups.