The purpose of this research is to develop a new method of digital subtraction angiography (DSA) that can be applied to real time with reducing motion artifacts caused by heart movement and respiration. To create the mask image for DSA, the maximum pixel value at each pixel (which is the opposite pixel value to that of a vessel filled by contrast medium) was selected from the previous 14 image frames. The search area for the maximum pixel value was selected using the value of the standard deviation (SD) for each pixel from the previous 14 image frames. When the SD value in the 14 frames was greater than a threshold level, the search area of the maximum value became 1 pixel x 1 pixel x 14 frames; otherwise, 7 pixels x 7 pixels x 7 frames. The image quality of new DSA was evaluated on 20 coronary arteriogram images, including various degrees of occlusion or stenosis. The results indicated a considerable improvement in DSA image quality; thus, the coronary arteries, carotid artery, and vein were clearly enhanced.