The pulsatile blood flow through human aortic arch and three major branches are computationally studied to investigate the effect of blood rheology on the hemodynamic parameters. The human aorta model is reconstructed from the computed tomography (CT) images of specific patient. The results of nine non-Newtonian (Casson, K-L, Modified Casson, Carreau, Carreau-Yasuda, Cross, Power-law, Modified Power-law, and Generalized Power-law) models are analyzed and compared with those of Newtonian model and reveal very interesting hemodynamic features for each model. Among the applied non-Newtonian models, the Cross model displays significantly different distribution of wall shear stress (WSS) and velocity field through the aorta at diastole. Comparing the local shear stress magnitude in three branches at different critical cardiac instants shows that the shear thinning nature of blood can slightly influence WSS at diastole, in all branches. The effect of blood rheology appears clearly in the brachiocephalic and carotid branches, at peak systole. In the high-shear rate zones, the lowest WSS is estimated by the Carreau model. The Newtonian model has close prediction to the Cross model at peak systole. The power law model predictions remain the nearest to those of the Carreau model along the cardiac cycle. (C) 2014 Elsevier B.V. All rights reserved.