Abstract: We propose a robust method for realizing stochastic predictive fault-tolerant switching control (SRPFTSC) for a class of industrial processes characterized by uncertainty, unknown disturbance, and random actuator failure. Our main contribution is the integration of the stochastic control theory with a robust predictive control method, which solves the problem of actuator failure in actual industrial processes and performs better than the traditional fault-tolerant control (FTC) method. First, we establish a new state-space extended model with multiple degrees of freedom, a system state increment, and output error tracking to describe industrial processes characterized by uncertainties, unknown disturbances, and random actuator failures. Secondly, based on the extended model, we design a feedback control law with fault stochastic characteristics to realize FTC with faults and normal control without faults. Correlation theorems and corollaries with linear matrix inequality are provided. To ensure that the system meets the requirement of H_∞ asymptotic stability, the stability condition of the switching of the fault-tolerant controller is obtained using an exponential stability proof. The effectiveness and feasibility of the method are verified by the results obtained from an injection molding process.