The performance of electrocardiogram applications is closely related to the succesful detection of R-peaks. Therefore, successful detection of R-peaks is of high importance. Electrocardiogram signals vary from person to person, so detection of R-peaks is getting difficult. In this study, Evolutionary Neural Network which is a deep learning model and imitates human vision, is used to detect R-peaks. The MIT-BIH Arrhythmia Database with 48 different records was used as the database and experimental results with very high accuracy were obtained.