The aim of this study is to establish the treatment couch model of TomoTherapy in a RayStation treatment-planning system. Point-dose measurements using a cylindrical solid phantom and ionizing chamber were performed to determine the optimal mass densities of the couch components by minimizing the discrepancy between the measured and calculated attenuations. To validate the efcacy of the determined mass densities, patient-specifc delivery quality assurance (DQA) of the TomoDirect plans for eight palliative spine radiotherapy (PSR), fve postmastectomy radiotherapy (PMRT), and TomoHelical plans for ten treatment sites were performed using an ionizing chamber, EBT3 flm, and Delta4 diode detector with the default couch model (DCM) and corrected couch model (CCM). The dose diferences and global gamma statistics at 3 mm/3% and at 2 mm/2% were evaluated and statistically compared between the DQA plans of DCM and CCM. When applying the default and optimal couch models for treatment planning, the mean dose diferences between the measured and calculated attenuations were − 4.8% and − 0.1%, respectively. In all DQA for PSR cases using the CCM, the dose diference and gamma statistics were statistically improved compared to those using the DCM (P<0.05). However, there was no signifcant diference in the dose diferences and gamma statistics between the PMRT and TomoHelical cases, except for the flm DQA for PMRT. The results demonstrated that the couch model in RayStation should be commissioned for TomoTherapy treatment planning when a greater dose contribution is expected from the posterior direction.