In this paper, an observer-based input-to-state stable approach is proposed to stabilize teleoperation systems over communication time-delay with better transparency compared with passivity approaches. In this paper, input-to-state stable approach is proposed for time-delayed teleoperation systems based on the hysteresis-like behavior of time-delayed communication network. Based on the investigated hysteresis behavior, input-to-state stable approach is applied on each transmitted signal to make the generated energy due to time-delay bounded. However, this approach could stabilize the teleoperation over a limited range of time-delay, and for larger time-delay, it could not guarantee stability of the time-delayed teleoperation. This paper also extends the main idea for teleoperation over larger communication time-delayed. For the communication network, hysteresis-like property of all the path in input vs. output graph of transmitted graph is extracted. Then, an input-to-state stable (ISS) observer is introduced to check the current condition of hysteresis property, and an adaptive control architecture is proposed to bound the generated energy based on the observed hysteresis condition. In the proposed approach, a finite amount of energy is allowed to pass through the system, therefore it provides less conservative control approach as well as higher transparency compared with passivity approaches. In addition, the proposed control approach is a unified framework for time-delayed teleoperation systems where it is able to dissipate the generated energy induced by time-delay regardless of selected bilateral control architecture. The extended ISS approach was implemented on a teleoperation system with communication time-delay, and the experimental results demonstrated that the proposed observer-based approach can stabilize the system over larger range of communication time-delay compared with the non-observer-based ISS approach.