Myocardial motion can be helpful in diagnosing the heart abnormalities because it relates to cardiac vascular supply. Analyzing the motion of all segments of the myocardial is a difficult task because of the random noise of echocardiographic images. Noise in the echocardiographic images that is produced by the image acquisition is not consistent in each frame. It causes error in motion computation especially in optical flow computation that relies on brightness constancy. To increase the accuracy of motion estimation along the myocardial boundary, we proposed a method that combines boundary detection and optical flow to compute myocardial motion at the myocardial boundary. In this method, computation of optical flow from 2 consecutive images is done after myocardial boundary is detected in each frame. The result shows that the motion estimation algorithm along the myocardial boundary yields better result compared to without using boundary extraction methods.