In order to confirm the validity of the ideal elasto-plastic resistance model applied to the ballast track under seismic loading, this paper studies the seismic response of continuous welded ballast track on bridges through the shaking table test and presents a process of updating the model based on the test results. The results indicate that the track constraint can improve the low order natural frequency of bridges significantly, and reduce the displacement response of the bridge. When ballast beds are effectively in a dynamic reciprocating state while under seismic loading, a structural change between the granules will occur, wherein some will flow and redistribute. The dynamic hysteretic change of the ballast longitudinal resistance is complex and quite different from that of the ideal elasto-plastic hysteretic route, and the ballast longitudinal resistance performance degenerates. If ballast longitudinal resistance is assumed to be ideal elastic-plastic resistance, the actual beam displacement response will be underestimated and the calculated rail seismic force will be greater than the test result. Moreover, the equivalent stiffness coefficient and damping coefficient of the ballast dynamic resistance characteristics could be obtained by model updating, and the simulation results coincide well with the test results.