Abstract: Objective To compare the intra-fractional setup errors between volumetric modulated arc therapy（VMAT） and static intensity modulated radiotherapy（IMRT） for cervical or upper thoracic esophageal carcinoma. Methods Fifty cervical or upper thoracic esophageal carcinoma patients were selected in this study. VMAT plans with two single arcs and IMRT with nine fields designed for each patients. Patients received cone-beam computed tomography（CBCT） scans before initial setup， after re-positioning and after radiation delivery. The CBCT images were registered to the planning CT images， then the intra-fractional setup errors were obtained and the differences were analyzed. Results In IMRT group， the after re-positioning error was（0.63±0.47）mm，（0.84±0.35）mm，（0.67±0.41）mm at X， Y， Z axes，the after radiation delivery was（1.01±0.42）mm，（1.08±0.89）mm，（1.07±0.70）mm，respectively. The latter was higher than the former， and the difference was statistically significant（P＜0.05）. In VMAT group， the after re-positioning error was（0.62±0.50）mm，（0.78±0.40）mm，（0.72±0.54）mm at X， Y， Z axes，the after radiation delivery was（0.71±0.52）mm，（0.84±0.41）mm，（0.79±0.63）mm， respectively. The latter was little higher than the former， and the difference was no statistical significance（P＞0.05）. The average error in X，Y，Z axes of both groups were increased， and the statistical datas of IMRT group were higher than the VMAT（P＜0.05）. The treatment time and monitor units in VMAT group were （2.85±0.73）min and 589.00±63.00, which was better then （8.14±1.06）min and 792.00±83.00 in IMRT group（P＜0.05）. Conclusion VMAT could greatly shorten the treatment time， reduce the influence of uncertain factors and patients discomfort， effectively reduce the intra-fractional setup errors. In addition， the error was gradually increased with the increasing of treatment time.