Although most hospitals have an emergency department disaster plan, most never have been implemented in a true disaster or been tested objectively. Computer simulation may be a useful tool to predict emergency department patient flow during a disaster. The aim of this study was to compare the accuracy of a computer simulation in predicting emergency department patient flow during a mass-casualty incident with that of a real-time, virtual, live exercise. History, physical examination findings, and laboratory results for 136 simulated patients were extracted from the disastermed.ca patient database as used as input into a computer simulation designed to represent the emergency department at the University of Alberta Hospital. The computer simulation was developed using a commercially available simulation software platform (2005, SimProcess, CACI Products, San Diego CA). Patient flow parameters were compared to a previous virtual, live exercise using the same data set. Although results between the computer simulation and the live exercise appear similar, they differ statistically with respect to many patient benchmarks. There was a marked difference between the triage codes assigned during the live exercise and those from the patient database; however, this alone did not account for the differences between the patient groups. It is likely that novel approaches to patient care developed by the live exercise group, which are difficult to model by computer software, contributed to differences between the groups. Computer simulation was useful, however, in predicting how small changes to emergency department structure, such as adding staff or patient care areas, can influence patient flow. Computer simulation is helpful in defining the effects of changes to a hospital disaster plan. However, it cannot fully replace participant exercises. Rather, computer simulation and live exercises are complementary, and both may be useful for disaster plan evaluation.