Platelet collagen receptor glycoprotein VI (pGPVI) is elevated in patients with acute coronary syndrome (ACS) and ischemic stroke. Recently, we developed a novel bead-based sandwich immunoassay to determine soluble GPVI (sGPVI), which has been validated in ACS patients. This study aimed to evaluate the plasma levels of sGPVI and pGPVI expression in patients with suspected stroke. We consecutively evaluated 176 patients, who were admitted to the stroke unit. Surface expression of pGPVI was determined by flow cytometry, sGPVI concentrations were determined using our sandwich immunoassay. Unlike patients with TIA, patients with stroke showed significantly decreased plasma levels of sGPVI compared to patients with non-ischemic (NI) events (TIA: mean [µg/L] ± standard deviation): 6.1 ± 2.1 vs. NI: 8 ± 4; p = 0.192; stroke: 5.9 ± 2.3 vs. NI; p = 0.013), whereas for pGPVI, patients with TIA and ischemic stroke revealed a significantly increased platelet surface expression compared to NI patients (TIA: mean fluorescence intensity [MFI] ± standard deviation): 20.9 ± 5.4 vs. NI: 17.6 ± 5.2; p = 0.021; stroke: 20.3 ± 6.2 vs. NI; p = 0.016). Using logistic regression analysis, both sGPVI (p = 0.002) and pGPVI (p = 0.012) are independently associated with ischemic stroke compared to other laboratory markers. To predict the individual risk for ischemic stroke using the plasma levels of sGPVI, receiver operating characteristic (ROC) analysis determined an optimal cutoff value of sGPVI at 6.5 µg/l, thus, patients with decreased plasma levels (<6.5 µg/l) have a 1.5-fold adjusted odds ratio (95%confidence interval, 1.4-2.7). Lower plasma levels of sGPVI are associated with the slightly elevated risk of stroke and may be a promising novel biomarker.