The finite and infinite horizon optimal control problem is studied for networked control systems wherein actuators are activated in groups by a stochastic event. The system is built as a switching Markov jump linear system according to the medium access mechanism of actuators. Based on the theory of Markovian jump linear system and dynamic programming, the optimal control sequence is derived to satisfy the quadratic performance index. The calculation technique of optimal control law is given by solving stabilizing solutions of coupled Riccati equations to achieve the stability of networked control systems in the exponentially mean square sense. Finally, a numerical example is given to demonstrate the effectiveness of the proposed method.