In this paper, a strategy for the reduction of dimensionality of nonlinear data based on radial basis function network and polygonal line algorithm is proposed. This strategy utilizes the polygonal line algorithm to define the number of nodes in the hidden layer of the network, which is mostly heuristic in case of other proposed methods. All the parameters related to the hidden layer are calculated using a polygonal line algorithm and hence reduce the training complexity. Kernel density estimation is used for robust estimation of the confidence limits. The proposed methodology is applied for fault detection and identification. A twin reactors virtual plant is selected as the case study to show the efficiency of the proposed strategy. The result shows that the proposed method is a promising direction for the fault detection and identification in real-time, nonlinear systems.