Abstract: Based on the finite element method, artificial neural network (NN) technique and reliability-based robust design theory, the explicit relational expression between the design parameters and the system reliability of the mechanical components with stress concentrations are given correctly by NN with the approximation concepts inherent in reliability analysis. On the basis of the trained NN, the probabilistic perturbation method can be used to obtain the reliability-based robust design information. Therefore, it solves the explicit expression of the limit state function that is difficult to obtain in engineering, and then the process of reliability-based robust design can be implemented conveniently.