This article explores the utility of neural network ensembles in knowledge discovery and integration. A novel neural network ensemble model KBNNE (Knowledge-Based Neural Network Ensembles) integrating KDD (Knowledge Discovery in Database) techniques and neural network modeling algorithms by "parallel operations" is proposed. Through balancing the relative importance of knowledge learned by induction and deduction, KBNNE can avoid the knowledge loss and enhance the "transparency " of neural network models. The effectiveness of the proposed model is demonstrated through computer simulations on simple artificial problems and an actual modeling problem.