<p>In image diagnosis of arteriosclerosis, it is important to observe continuous structure of space by 3D information in order to prevent misdiagnosis and oversight. A three-dimensional ultrasonic-measurement-integrated (3D-UMI) blood flow analysis reproduces in vivo blood flow field by feeding back the continuous cross-sectional information of the blood flow Doppler velocity obtained by scanning the 1D ultrasonic probe to the numerical fluid analysis. However, an effect of the inflow condition on analysis accuracy is known in 2D-UMI analysis with 1D linear probe, and this effect for 3D-UMI analysis with 2D linear array probe is unknown. The purpose of this study was to clarify the spatial characteristic of analysis accuracy in 3D-UMI blood flow analysis with 2D linear array probe by a numerical experiment. A standard solution was calculated as a blood flow model of a common carotid artery, and 3D-UMI analysis were performed for the Doppler velocity of the standard solution. As a result, it was clarified that the reproducibility of the slow flow field is relatively good, but that of the fast one is not good enough because a bias of the error distribution is caused by the feedback boundary.</p>