In this paper, an Extended Kalman Filter (EKF) and a model-dependent nonlinear controller over network using the separation principle for Low Earth Orbit (LEO) satellite Attitude Determination and Control Subsystem (ADCS) have been designed. In this context, according to the satellites development trend, ADCS architecture for a broad class of LEO satellites is proposed to stabilize and achieve mission objectives such as precision attitude determination and pointing. This architecture is a Networked Control System (NCS) used to establish connection and communication among control components including sensors, actuators and onboard processors, as well as to share data with other subsystems. Then, by modeling all components of the system, and considering the network effects as a bounded disturbance, the control system is designed to compensate of these effects. For this purpose, estimation and control algorithms including EKF and a model-dependent nonlinear controller is designed such that in addition to achieve desired system performance, the stability of each of them is guaranteed. Afterwards, the nonlinear dynamics model of the satellite in terms of quaternion parameters and angular velocities is presented, and by expression of the separation principle for nonlinear observer and controller design, their convergence and exponential stability conditions based on linearized model of satellite are derived. Proof of theorem shows that the closed-loop system continuously maintained satellite attitude in the specified accuracy range. Finally, simulation results obtained from applying the designed observer and controller on the active satellite in orbit demonstrates the efficiency of the proposed design.