Abstract: Genetic Algorithms is Optimal Algorithm, which employ a search technique based on ideas from Darwin's natural evolution theory. A new genetic algorithm (NGA) is proposed in this paper, which not only can keep the population diversity but also has quicker convergence speed. The experiment results show that the improved genetic algorithm can efficiently find global optimal beyond premature convergence. Finally, using the NGA, optimizing the topology and training the weights for BP neural network are done. The results of the applications show that the new genetic algorithm is practical and efficient.