Nonlinear systems sometimes suffer from modeling errors which could potentially result in instability and reduce the control performance of the actual systems. This paper reviews some previous research studies examining the sources of these errors and how to deal with them. In this work, a fuzzy Lyapunov function is defined through fuzzy blending quadratic Lyapunov functions to avoid conservatism in the stability conditions. Therefore, the neural network (NN) model based approach is provided and the fuzzy systems can be transferred into a linear difference inclusion (LDI) formulation. Time-delay stability conditions of closed-loop controlled systems are then derived based on robustness design to deal with the modeling error problems. Key words: Fuzzy Lyapunov method, LDI, T-S fuzzy systems, artificial intelligence.