Handwriting-gesture recognition has been widely implemented in handwriting input application. Usually, gestures are used to conduct edit operations or be set as short-cut of an application. In this paper, we compare several handwriting-gesture recognition methods, and address their different user cases. These methods include pixel-matching method, rule based method and discriminant-function based method. For discriminant-function based method, we describe 2 sub-methods. They are prototypes based method and training based method. We not only analyze recognition accuracy of gestures for these methods, but also analyze their distinguished capability when recognizing gestures and alphanumeric in same recognizing mode. Experiments results show that, if the gesture-samples are enough, training based method achieves the highest accuracy. Furthermore, when recognizing mixed input of gestures and other handwriting symbols, training based method almost doesn’t degrade accuracy of these symbols.