[Purposes] Due to aerodynamic interference between parallel bridges, the installation of wind barriers significantly increases the characteristic height of the bridge, and the leeward side bridge is prone to wake-induced vibration. To ensure the accuracy of the experiment and the safety of the project, it is necessary to study the size effect of wake-induced vibration response of parallel bridges. [Methods] Wind tunnel tests were conducted on parallel cable-stayed bridges with girder segment models at scale ratio of 1:20 and 1:50, respectively. The effects of damping ratio, wind attack angle, and girder shape on the size effect were discussed. [Findings] The results show that both girders of the two bridges have significant size effects. Among them, the normal scale model test of the railway bridge has a higher peak amplitude compared to the large scale model test, and the wind speed range is basically similar. The damping ratio and wind attack angle have a relatively smaller impact on the scale effect of the wake-induced vibration response. For the same streamlined closed box girder, differences in sectional dimensions can also have a certain impact on the scale effect. For flat cross-section highway bridges, vertical and torsional vibration lock-in regions are wider in the normal-scale model test, with larger amplitudes. At a -3° wind attack angle, due to the diversion effect of the railway bridge wind barrier and the buffet of the normal scale model, the peak vertical and torsional vibration amplitudes in the normal-scale model test were lower than those in the large-scale model test, contrary to the results under other conditions. [Conclusion] The size effect of wake-induced vibration varies depending on the cross-sectional dimensions of the same streamlined closed box girder.