The Wiener Functional Series are used to describe the analog circuit of an undiscovered nonlinear characteristic and the circuit character is represented as the value of Wiener kernel. However, the complexity of the Value of Wiener kernel extraction is growing at an exponential rata with the increases of orders and sampling points and it is difficult for application. In view of this problem, a folded and recursive fast algorithm was proposed in this paper and over90% of the computational cost can be reduced with the fast algorithm than primary algorithm based on the careful analysis and calculation. On this basis, the experiments were conducted that compared the results of primary algorithm and fast algorithm applying on typical circuit and the conditions of circuit among normal and different soft faults were rapidly and efficiently determined.