A new switching control strategy is put forward for time-varying delay batch processes with actuator faults satisfying some probability. The batch process is first expressed as a 2D stochastic delayed system and an iterative learning control (ILC) law is designed. Then an equivalent 2D stochastic delay switched system is constructed to design the ILC law. Based on stochastic Lyapunov stability theory, stochastic stability conditions of the above equivalent system depending on the upper and lower bounds of delay are offered, the update law under different fault probabilities is designed, and the new switching control strategy-based average dwell time (ADT) method under both normal and fault situations is provided. The merits of the proposed switching control strategy are to ensure that batch processes under different conditions (normal or fail) achieve tracking control and to lower the operation risk, together with energy savings suggestion. Finally, control of packing pressure in an injection molding process is taken as an example to verify the effectiveness and practical value of the proposed method.