Abstract: The SVM (Signal Vector Magnitude) values used in most behavior recognition studies are appropriate for recognizing dynamic behavior but difficult to recognize static behavior. Therefore, we introduced “Polar angle” to recognize static behavior. This is the method of using the angle between the measuring sensor and the axis of gravity acceleration opposite direction. In this study, we propose a method using SVM and angle to effectively recognize dynamic behavior and static behavior. Experimental results show that the proposed method achieves an average of 95.55% accuracy.