Background Patients with a sudden onset of focal neurological deficits consistent with stroke, who turn out to have alternative conditions, have been labeled stroke mimics. Aims We assessed a recently validated telemedicine-based stroke mimic score (TeleStroke mimic score; TM-score) and individual patient characteristics with regard to its discriminative value between cerebrovascular disease and stroke mimic patients in the in-person, pre-hospital setting. Methods We evaluated patients cared for in a mobile stroke unit in Berlin, Germany. We investigated whether the TM-score (comprising six parameters), Face Arm Speech Time test, and individual patient characteristics were able to differentiate cerebrovascular disease from stroke mimic patients. Results We included 423 patients (299 (70.7%) cerebrovascular disease and 124 (29.3%) stroke mimic) in the final analysis. A TM-score > 30 indicated a high probability of a cerebrovascular disease and a score ≤15 of a stroke mimic. The TM-score performed well to identify stroke mimics (area under the curve of 0.74 under receiver-operating characteristic curve analysis). The cerebrovascular disease patients were older (74.8 vs. 69.8 years, p = 0.001), had more often severe strokes (NIHSS > 14 25.8% vs. 11.3%, p = 0.001), presented more often with weakness of the face (70.9% vs. 42.7%, p = 0.001) or arm (60.9% vs. 33.9%, p = 0.001), dysarthria (59.5% vs. 40.3%, p < 0.001), history of atrial fibrillation (38.1% vs. 21.0%, p = 0.001), arterial hypertension (78.9% vs. 53.2%, p < 0.001), and less often with seizure (0.7% vs. 21.0%, p < 0.001). Conclusions The TM-score and certain patient characteristics can help paramedics and emergency physicians in the field to identify stroke mimic patients and select the most appropriate hospital destination.