Combining fault diagnosis with fault tolerance control can reduce the operating cost of the system. A neural network recognition method is established to model the dynamic behavior of the controlled system. The scale coefficient parameters and control law of the system are optimized. A method based on two-step optimization is proposed to solve this problem. By analyzing the redundancy of each component, the fault tolerance of the system after failure is studied, and the fault tolerance of the regulator is quantified. By studying the influence of each component on the static and dynamic performance of the control system, the main components leading to the failure of the control system are found out. Through computer simulation, the feasibility of this algorithm in thermal engineering system is verified. The fault diagnosis and fault-tolerant control system are designed for the key equipment of a certain nuclear power plant, and the two control strategies are simulated. The research results show that the application of this system has certain application value for the fault diagnosis of the key equipment of nuclear power plant, and can provide guarantee for the safe and stable operation of nuclear power plant.