An accurate prediction of clinical outcomes in acute ischemic stroke is essential for improving patient management. There is growing research interest in EEG-based neurophysiological biomarkers. In this context, although the prevalence of epileptiform patterns is emerging, their role as potential predictors of outcome is still poorly investigated. In this study, we investigate whether interictal epileptiform discharges (IEDs) detected in the acute phase can contribute to predicting long-term functional outcomes in treated ischemic stroke patients and to the development of a predictive model. We retrospectively analyzed clinical, neurophysiological, and neuroimaging data from 228 acute ischemic stroke patients (mean age: 74 ± 14 years; 122 females). Patients were classified into two groups based on their 3-month modified Rankin Scale (mRS) scores: good outcomes (mRS 0–2, 127 patients) and poor outcomes (mRS 3–6, 101 patients). Standard 19-channel EEG recordings were conducted to assess epileptiform abnormalities and IEDs. Feature selection was performed based on the Gain Ratio method, and predictive models were developed using five different machine learning (ML) techniques: classification tree, logistic regression, naïve Bayes, artificial neural network, and support vector machine. Our results identified the National Institutes of Health Stroke Scale (NIHSS) at admission, mRS anamnestic, chronic heart failure, acute infectious disease, sum of lobes with a infarct area, atrial fibrillation, and IEDs as significant predictors of ischemic stroke functional outcomes. Logistic regression achieved the best performance, with an accuracy of 81% and an AUC of 0.83. These findings confirm the importance of clinical and radiological features while also highlighting the predictive value of interictal EEG epileptiform patterns for stroke outcome prediction.