Background and Purpose— The perfusion territories of the brain-feeding arteries are difficult to assess in vivo and therefore standard cerebral perfusion territory templates are often used to determine the relation between cerebral infarcts and the feeding vasculature. In the present study, we compared this infarct classification, using standard templates, with the individualized depiction of cerebral perfusion territories on MRI. Methods— The ethics committee of our institution approved the study protocol. A total of 159 patients (92 male, 67 female; mean age, 58.9 years) with first-time clinical symptoms of cerebral ischemia were included in the study. Diffusion-weighted imaging was used for depiction of the area of ischemia and the perfusion territories of the left internal carotid artery, right internal carotid artery, and vertebrobasilar arteries were visualized with territorial arterial spin labeling MRI. Infarct locations with respect to cerebral perfusion territories were evaluated with and without territorial arterial spin labeling MRI images. Results— In 92% of the patients, the territorial arterial spin labeling images were of diagnostic quality. One hundred thirty-six patients showed areas of ischemia on diffusion-weighted images. The additional information from the territorial arterial spin labeling images changed the classification in 11% of the cortical or border zone infarcts (6 of 56), whereas no territorial changes were observed in lacunar, periventricular, cerebellar, and brainstem infarcts. Conclusion— The diagnostic information provided by perfusion territory imaging in patients with stroke is valuable for the classification of cortical and border zone infarcts, whereas no change of the textbook-based classification was observed for other infarct types.