Atrial fibrillation (AF) is a relatively frequent complication of acute myocardial infarction (AMI), with a reported prevalence of up to 23% and an independent association with worse clinical outcomes. Identifying patients at high risk of AF during a hospitalization for AMI could allow to conceive more targeted monitoring and preventative measures. While AF prediction has been studied in several patient populations, the identification of risk factors for early, new-onset AF (NOAF) after AMI in the intensive cardiac care unit (ICCU) context remains less explored. In particular, to our knowledge, there are no reported attempts at predicting in-hospital NOAF after AMI using machine learning. In this retrospective study, we aimed at developing a machine learning model to predict in-hospital NOAF following AMI. The clinical dataset used for model development included 2445 consecutive AMI patients admitted to the ICCU of Centro Cardiologico Monzino, out of which 241 (9.9%) developed NOAF prior to ICCU discharge. Fifty-six features encompassing demographic and clinical variables (clinical history, biochemical data, therapy at admission, etc.) were retrospectively collected and analysed. To select the best prediction model, a set of data balancing, feature selection and classification techniques were evaluated and compared by means of area under the ROC curve (AUROC) through nested cross-validation. The best-performing model, combining a feature selection exploiting mutual information and a logistic regression classifier, achieved an AUROC of 0.754 (95% CI: 0.721 – 0.786). Importantly, the feature selection process of our model retained both known risk factors and previously unreported markers. These findings provide new insight on risk assessment for early NOAF after AMI, contributing to a growing knowledge on this topic.