Background: The accuracy of radiotherapy is based on the matching of 2D portal/CBCT image with a reference image. The aim of this study is to determine the random and systematic setup errors (in cm) in radiotherapy of head and neck cancer patients and to derive the setup margin and its clinical implications.Methods: Author retrospectively reviewed the records of 25 head and neck cancer (HNC) patients treated with radiotherapy between Dec 2017 and July 2018. After immobilization, setup accuracy was assessed by registration of XVI image with planning reference image using Elekta XVI image guidance system and the isocenter correction was applied. For each patient 10 CBCT image sets were taken. The translational errors in X, Y and Z directions were used to estimate systematic (Σ) and random (σ) errors and to derive the final setup margin by using van Herk’s formula (2.5Σ + 0.7σ).Results: The mean translational errors ranges from -0.23 cm to 0.32 cm in Lateral (X), -0.15 to 0.16 cm in Longitudinal (Y) and -0.11 to 0.17 cm in vertical (Z) directions. The Mean and SD for systematic errors 0.21±0.13, 0.11±0.18, 0.14±0.11 and random error (in cm) are -0.03±0.33, 0.00±0.21 and 0.05±0.30 in X, Y and Z axis respectively. The final total margin for CTV to PTV including setup margin in the X, Y and Z directions (in cm) were 0.56, 0.61, and 0.47 respectively.Conclusion: Thus, the precise immobilization techniques are very important to reduce the setup margins, and the number of CBCTs during head and neck radiotherapy treatment.