The stabilization of the man-made dynamic systems has been achieved by sensor based state feedback control algorithms which require high computational bandwidth and high stiffness structures. However, many biological systems achieved similar or superior stable behavior with low speed signal transmission via nervous systems, which is easy to introduce unstable performance in the view of control engineering. In order to explain this phenomenon, the concept of self-stabilization has been recently proposed and investigated widely. Self-stabilization is defined as the ability to restore its original state after a disturbance without any feedback control. We analytically investigated the stabilizing function of a musculoskeletal system using the Lyapunov stability theory. Based on this investigation, in this study, we propose a design method to realize the self-stabilizing function of a musculoskeletal system, and experimentally verify that the self stabilizing function can be physically realized and explained by the proposed Lyapunov function.