Complex fractionated atrial electrograms (CFAEs) have been reported as effective targets for ablation in the treatment of atrial fibrillation. CFAEs are defined as atrial electrograms with a very short cycle length (<120ms) or continuous electrical activities. However, methods for detecting CFAE automatically and quantitatively using intracardiac electrograms have not yet been established. The purpose of this study is to propose a method to automatically detect CFAE by detecting atrial electrograms that have either a short cycle length (<120ms) or continuous electrical activities, as well as to confirm the effectiveness of the method using clinical data. The measured electrograms were squared in order to amplify electric potential differences between the atrial electrograms and continuous electrical activities. This method detects CFAE by counting the numbers of times the squared electrogram exceeds one of two thresholds. The first threshold for detecting atrial electrograms with a short cycle length (<120ms) is generated from the variance of the electrograms. The second threshold for detecting continuous electrical activities is obtained from the maximum value of atrial electrograms multiplied by a reduction factor. We analyzed 100 electrograms consisting of 62 CFAEs and 38 non-CFAEs, measured from eight persistent atrial fibrillation patients, with durations ranging from 4 to 32 s. It was confirmed that the proposed method could automatically detect CFAE and non-CFAE with 93% accuracy (positive predictive value: 97%, negative predictive value: 88%). In conclusion, the result of this study indicates that the proposed method is effective for automatic CFAEs detection. We found that it is feasible to use the proposed method for automated CFAEs detection without the need for an operator's judgments during the treatment of persistent atrial fibrillation. Further studies are warranted to determine the clinical significance of the proposed method.