Wind tunnel tests are carried out on an overpass railway bridge to investigate the vortex-induced vibration (VIV) performance of the main beam section with anti-throwing wall. The test results show that compared with the main beam section with anti-throwing net, the VIV phenomenon is obvious as well for this section. Moreover, the VIV performance is worse. In order to suppress the VIV, several aerodynamic optimization measures are discussed, including setting the flow guide plate, moving the track of the repair car, sealing the anti-collision railings, setting the upper central stability plate with whole sealing or air permeability, setting the flow suppression plate and so on. It is found that the upper central stability plate and the flow suppression plate with inward tilt are effective measures for the suppression of VIV. As for the upper central stability plate, the suppression effect becomes better for larger height. In addition, the effect is worse under the increase of the air permeability. When the width of the flow suppression plate and the height of the upper central stability plate are larger simultaneously, the effect of their combination is prior to that of either of them. The results can provide reference for wind-resistant design of this kind of bridge.