The stick-slip phenomenon is used to explain as a mechanism of earthquake recurrence. A number of stick-slip experiments have been performed to clarify the mechanism of recurring slip instabilities and slip weakening. Although the amplitude of sliding of most experiments is quite smaller than actual earthquakes, and the observed acceleration is larger. The authors have developed a new experimental setup, in which blocks move on a conveyor belt and is restrained by the spring, and conducted stick-slip experiments. This experimental setup is able to simulate conditions in actual earthquakes better than previous stick-slip experimental devices. In this study, we conducted stick-slip experiment which focused on difference of the rock types and confirmed the the applicability of this experimental setup and the validity of the result. Moreover, the theoretical results of stick-slip phenomenon indicate the need for taking account of slip weakening and roughness of the contact surface and geometrical shape.