Purpose. Ischemic or hemorrhagic stroke are among the most common causes of seizures, especially in the elderly. EEG is the only technique that can detect epileptiform abnormalities (EA) and nonconvulsive status epilepticus (NCSE), which may negatively affect recovery of these patients. Herein we aimed to investigate the potential predictive value of long-term EEG findings in terms of poststroke seizures (PSS) and survival, in stroke patients followed up in the neurological intensive care unit (NICU). Methods. Video-EEG reports of stroke patients (ischemic and hemorrhagic) hospitalized between 2009 and 2014 in our NICU were reviewed. Patients with <2 months of survival were excluded. Follow-up data were obtained via telephone calls or patient charts. The correlation between EEG findings and early (<= 1 week) and late seizure (>1 week) occurrence, recurrent seizure development, outcome, and survival were analyzed statistically. Results. Overall 50 patients (27 female, 23 male; age, 26-85 years) were included in the final analysis. Almost 60% developed PSS (similar to 2/3 were early). There was no difference between ischemic versus hemorrhagic stroke patients (P = .72). Recurrent seizures were more common in the late seizure group (P < .001). EAs occurred in one-third of the study cohort. This finding did not predict seizure development (P = .93) or survival (P = .61). Conclusion. PSS are a frequent finding in stroke patients followed up in NICU. EAs are not uncommon, but do not predict seizure occurrence or survival.