A stochastic robust predictive fault-tolerant control approach is proposed for a class of industrial processes with uncertainties, external disturbances and the random failure of actuators. First, a process model of the above system is established in terms of state space. It is further converted into a new augmented form containing the state deviation and output tracking error. By means of the augmented model, the designed control law is given. Then the related theorem and corollary in the form of linear matrix inequality constraints are provided for solving the gain of the control law under different probabilities of failure. Under a certain probability, the occurrence of a failure is regarded as a switching condition. Conventional control is adopted when the failure does not occur, while fault-tolerant control is performed when the failure occurs. Finally, a pressure holding phase case is provided for verifying the feasibility of the given approach.