This paper concerns the fault tolerant control for discrete networked control systems (NCSs) with probabilistic sensor and actuator fault, random delay and packet dropout. The fault of each sensor or actuator happens in a random way, which is described by an individual random variable satisfying a certain probabilistic distribution. Using these stochastic variables in the system model, new type of NCSs fault model is proposed. The merit of the proposed fault model lies in its generalization and reality, which can cover some existing fault models as special cases. By using Lyapunov functional method and linear matrix inequality technology, sufficient conditions for the mean square stable (MSS) of the NCSs can be obtained. Then the reliable controller can be designed. Finally, a numerical example and a practical example are given to demonstrate the efficiency and application of the proposed method.