Diagnosing delirium in neurointensive care is difficult because symptoms of delirium, such as inappropriate speech, may be related to aphasia due to primary brain injury. Therefore, validated screening tools are needed. The aim of this study was to compare two Czech versions of already validated screening tools the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU) and the Intensive Care Delirium Screening Checklist (ICDSC) - in a cohort of acute stroke patients. We also aimed to assess the pitfalls of delirium detection in the context of non-convulsive status epilepticus (NCSE). We analysed 138 stroke patients admitted to the neurological intensive care unit (ICU) or stroke unit. According to expert judgement, which was used as the gold standard, 38 patients (27.54%) developed delirium. The sensitivity and specificity of the ICDSC were 91.60% and 95.33%, respectively, and the positive and negative predictive values were 76.76% and 98.54%, respectively. Similarly, the sensitivity and specificity of CAM-ICU were 75.63% and 96.74%, respectively, and the positive and negative predictive values were 79.65% and 95.93%, respectively. We did not detect an episode of NCSE mimicking delirium in any of our stroke patients who were judged to be delirious by expert assessment. Our results suggest that the ICDSC may be a more suitable tool for delirium screening than the CAM-ICU in patients with neurological deficit. NCSE as a mimic of delirium seems to be less common in the acute phase of stroke than previously reported.