Introduction Postoperative myocardial revascularization atrial fibrillation (POAF) is a clinical complication that affects about 30% of patients and its mechanisms of origin are still poorly understood. This fact makes it difficult to identify the patient at greatest risk for this arrhythmia. This mission seems evident due to the complications it entails, including longer hospital stays, risk of stroke, heart failure, and death. There are reports of preoperative clinical aspects inherent to the patient’s condition, such as gender and age, and discontinuation of beta-blockers as risk factors. In addition, additional information obtained by electrocardiogram, echocardiogram, and blood count data, for example, present only modest predictive results. The analysis of heart rate and heart rate variability obtained by the Stroke Risk Analysis System (SRA) is a technique used to predict ambulatory atrial fibrillation (AF), using recordings of only one hour showing good accuracy. This system, however, has not yet been used to predict the emergence of POAF. The rationale for its use is based on the suspicion that the emergence of POAF is strongly related to sympatho-vagal imbalance and the increase in atrial ectopia, that is, changes in heart rhythm, the main variables analyzed by the SRA algorithm. Objective To assess the accuracy of the SRA to identify patients at risk of having POAF after coronary artery bypass graft surgery (CABG). Method 114 consecutive patients with coronary artery disease underwent coronary artery bypass grafting between the years 2015 and 2018. Between the first and fifth postoperative days, they underwent continuous electrocardiographic monitoring using the Holter system for cardiac rhythm analysis. Patients were divided into two groups: Group I was formed of those with POAF and Group II included patients without POAF. The tracings obtained by Holter were reanalyzed using the CardioManager®/Cardios program, converted and divided into one-hour sections using the SRA®/Cardios and Geratherm Converter program and submitted to the SRA-Apoplex medical/Geratherm® analysis algorithm. The SRA identifies three possibilities for classifying patient risk: a) Risk 0: patient in sinus rhythm; b) Risk 1: patient at increased risk for paroxysmal AF; c) Risk 2: patient with AF already present. For Group I, SRA were considered positive when Risks 1 and 2 were identified. For Group II, those identified as Risk 0 were considered negative SRA. Results POAF occurred in 33/114 patients (28%). The sensitivity, specificity, positive predictive value, and negative predictive value of the SRA to identify patients with POAF were 69%, 84%, 69%, and 82%, respectively; the positive and negative likelihood ratios, in addition to the accuracy of the SRA were, respectively, 4.3%, 0.36%, and 79%. A subanalysis of the results of the day on which AF occurred was performed on the records obtained in the first three hours of recording and up to three hours before the appearance of POAF. In the first period, the SRA was able to predict POAF in 57% of cases, while in the second period, the system identified the arrhythmia in 83% of cases. Conclusions a) The SRA presents good accuracy to predict POAF; b) its accuracy is moderate in the first three hours of recording; c) the accuracy increases significantly near the beginning of POAF; d) these findings indicate that electrophysiological changes that precede POAF are acute, occurring a few hours before the event and are identified by the SRA algorithm.