This article provides a quantitative error analysis of a simulation model of xenon/CT CBF in order to investigate the behavior and effect of different types of errors such as CT noise, motion artifacts, lower percentage of xenon supply, lower tissue enhancements, etc. A mathematical model is built to simulate these errors. By adjusting the initial parameters of the simulation model, we can scale the Gaussian noise, control the percentage of xenon supply, and change the tissue enhancement with different kVp settings. The motion artifact will be treated separately by geometrically shifting the sequential CT images. The input function is chosen from an end-tidal xenon curve of a practical study. Four kinds of cerebral blood flow, 10, 20, 50, and 80 cc/100 g/min, are examined under different error environments and the corresponding CT images are generated following the currently popular timing protocol. The simulated studies will be fed to a regular xenon/CT CBF system for calculation and evaluation. A quantitative comparison is given to reveal the behavior and effect of individual error resources. Mixed error testing is also provided to inspect the combination effect of errors. The experiment shows that CT noise is still a major error resource. The motion artifact affects the CBF results more geometrically than quantitatively. Lower xenon supply has a lesser effect on the results, but will reduce the signal/noise ratio. The lower xenon enhancement will lower the flow values in all areas of brain.