Stroke is a worldwide medical emergency and an important issue in stroke research is looking for the early pathophysiological markers which can predict the severity of brain injury. Decreased cerebral blood flow (CBF) has been serving as the most important indicator of ischemic stroke. Particular attention is paid to study the spatio-temporal CBF changes immediately after the onset of stroke in a rat intraluminal filament middle cerebral artery occlusion (MCAO) model and investigation of its correlation with brain infarct volume after 24 h. We implement an on-line laser speckle imaging (LSI) system, which could provide real time high spatio-temporal resolution CBF information before, during, and immediately after the rat MCAO surgery. We found a significant correlation between the affected area with 50% CBF reduction () at the first minute after occlusion with the infarct volume. To the best of our knowledge, this is the earliest CBF marker for infarct volume prediction. Based on such a CBF–infarct volume correlation, LSI may be used as a real time guidance for improving the consistency of intraluminal filament MCAO model since the depth of filament insertion could be adjusted promptly and those unsuccessful models could be excluded in the earliest stage.