Objective: To evaluate the impact of couch and collimator angular variations on dose volume histogram (DVH), tumour control probability (TCP), and normal tissue complication probability (NTCP) of the volumetric-modulated arc therapy (VMAT) plans. Methods: Stereotactic radiosurgery and stereotactic body radiation therapy VMAT plans were generated for three different hypothetical planning target volumes (PTVs) that mimic brain metastases, single brain lesion, and single spine lesion. Thirty routine VMAT plans (10 prostate, 10 head and neck, and 10 brain cases) treated in our clinic were also selected for this study. The plans were generated using an Eclipse Treatment Planning System and delivered using a Clinac iX linear accelerator equipped with a Millennium 120 multileaf collimator. All the plans were generated using two complementary full arcs (with gantry angle from 179 to 181 and collimator rotation of 30 and 330) except the brain tumour cases, which used single full arc with collimator rotation of 30. In all the cases, the couch angle was zero. Impact of the angular variations in the collimator and couch was studied by varying the collimator and couch angular settings by 1 , 2 , and 3 , and creating six erroneous plans corresponding to the original plans. The variation due to these errors on different DVH and radiobiological parameters (TCP, equivalent uniform dose (EUD), and NTCP) of the PTVs and organs-at-risk (OARs) were observed. The relative percentage of difference in these parameters (DD, DTCP, DEUD, and DNTCP) were analysed, and statistical significance was tested. Results: The variation due to collimator misplacement was observed to be larger than the couch misplacement. Furthermore, in both cases, the variation increased as the degree of error increased. Among the DVH parameters, D 98% , D 95% , and V 95Gy were affected more by the errors than D 2% , D 5% , and D 50% , in both hypothetical and clinical PTVs. In the clinical PTVs, the TCP showed the most variation among all parameters. The DNTCP of the bladder and brain OARs were zero, whereas for head and neck OARs, it was high.