An ANN-based dynamic FQFD method is proposed in order to solve the dynamic and fuzzy nature of QFD, that solves the problem of the dynamic nature by using neural network method, while the introduction of trapezoidal fuzzy number for its ambiguity. Firstly, a combined method with neural networks and FQFD is established, after learning and training, the method can quickly and effectively deliver customer requirements to the products designers. Finally, a case study demonstrates the effectiveness of the proposed method during the course of new products design.