We aimed to determine the effect of an AED with real-time feedback on the quality of bystander CPR. A before and after simulation study was designed. Trainees of BLS education were recruited for the simulation trial. Each team consisted of 2 participants with different roles (initial witness and CPR supporter). 82 teams performed the simulation scenarios with the real-time feedback function of the AED disabled initially, and then repeated it with feedback enabled. Quality measures of chest compression depth and no-flow fraction were compared between each of the two simulation scenarios. The mean percent of chest compressions with adequate depth was significantly higher in simulations with real-time feedback enabled (78.1% vs. 89.3%, p<0.01). In a subgroup with the lowest percent of adequate depth performance in the initial simulation without real-time feedback, a mean increase of 47.7% (95% CI 31.3-64.0) in the percent of chest compressions with adequate depth was seen with realtime feedback. Use of an AED with real-time feedback improves the quality of bystander CPR by laypersons in simulation of OHCA. The positive effect of real-time feedback is greatest among people with a low level of CPR skill.