Atherosclerosis is the main cause of circulatory diseases, and it is very important to diagnose atherosclerosis in its early stage. In an early stage of atherosclerosis, the luminal surface of an arterial wall becomes rough due to injury of the endothelium. Conventional ultrasonic diagnostic equipments cannot detect such micron-order surface roughness because their spatial resolution is, at most, 100 μm. In this study, for the accurate detection of surface roughness, ultrasonic beams were insonified from various angles relative to the surface of an object that has a micron-order asperity. Then, we focused on the angular dependence of echo amplitude and frequency characteristics. Using this method, it is shown that the angular dependence and frequency characteristics vary when an object has a surface roughness that cannot be detected by conventional B-mode imaging using linear scanning.