OBJECTIVE A small epidural hematoma (EDH) that has been diagnosed to be nonsurgical by initial brain computed tomography (CT) can increase in size and need surgical removal, resulting in a poor prognosis. However, there have been few studies, which focused delayed operated EDH. Therefore, we analyzed the clinical factors to determine the predicting factors of delayed operated EDH. METHODS Between January 2011 and January 2014, 90 patients, who were admitted due to EDH, were enrolled in this study. None of the patients were indicated for operation initially. Based on the presence of surgery, we classified the patients into a delayed-surgery group (DG) and a non-surgical group (NG). Additionally, we analyzed them according to the following: time interval between the trauma and the initial CT, gender, age, medical history, drinking, change of mean arterial pressure (MAP), volume of EDH and other traumatic brain lesion. RESULTS Among the 90 patients, the DG was 19 patients. Compared with NG, the DG revealed increased MAP, less presence of drinking, and a short time interval (DG vs. NG: +9.684 mm Hg vs. -0.428 mm Hg, 5.26% vs. 29.58%, 1.802 hours vs. 5.707 hours, respectively, p<0.05). Analyzing the time interval with receiver operating characteristic, there was 88.2% sensitivity and 68.3% specificity at the 2.05-hour cut-off value (area under the curve=0.854). CONCLUSIONS According to our results, the time interval between the trauma and the initial CT along with blood pressure change are potential predicting factors in the cases of delayed operation of EDH.