High-speed camera investigation of rapidly moving red blood cells (RBCs) in the microvasculature has been limited by an inability to handle the vast volume of data. We have developed a novel method to analyze large numbers of RBC images captured by a high-resolution, high-speed camera fitted on a confocal fluorescence microscope, to determine the velocities of individual RBCs in capillaries in vivo. Fluorescein isothiocyanate (FITC)-labeled RBCs flowing in the microvasculature of the cerebral cortex of urethane-anesthetized Wistar rats were recorded through the skull window on video clips during specified periods at high frame rates (500 fps). Sequential frames of moving RBCs in the video clips for a specified period were analyzed offline with in-house software (KEIO-IS2). Images of RBCs acquired were numbered automatically in order of appearance and displayed in a two-dimensional (2-D) RBC tracking map. The velocities of individual RBCs were automatically computed based on the RBC displacement per frame multiplied by the frame rate (fps), and the results were displayed in a 2-D velocity map and a 2-D RBC number map. Single capillaries were identified by staining with FITC-dextran. The mean capillary velocity of RBCs was evaluated as 2.05 +/- 1.59 mm/second in video clips obtained at 500 fps. This method is considered to have wide potential applicability.