Objective: To investigate flow patterns at carotid bifurcation in vivo with combining computational fluid dynamics (CFD) and MR A. Methods: Fifty patients with 100 normal carotid bifurcations proved with DSA or carotid ultrasound underwent contrast-enhanced MRA (CEMRA) of carotid artery. The source images were imported into CFD software to calculate and visualize flow pattern after post-processing and meshed. Flow patterns and eddy shape of bifurcation plane in reduced ejection phase of a cardiac cycle were analyzed. The carotid bifurcation angles were measured with vmtk software. Results: According to the shape of eddy current, the flow patterns could be classified as non-eddy type (n=8), one eddy type (n=25), two-eddy type (n=61) and multiple-eddy type (n=6). There were four types in bifurcation geometry based on the bifurcation angle and diameter of bifurcation, i.e. small bulb with small bifurcation angle (n=14), small bulb with big bifurcation angle (n=5), big bulb with small bifurcation angle (n=49) and big bulb with big bifurcation angle (n=32). The more big bifurcation angle and bulb, the more complicated flow pattern was observed at carotid bifurcation. Conclusion: CFD combined with MR angiography can be utilized to visualize and analyze flow patterns and the shape of eddy current of carotid bifurcation, therefore being helpful to understanding the relationship between hemodynamic factors and artherosclerosis. Copyright © 2013 by the Press of Chinese Journal of Medical Imaging Technology.