The author constructed the fitness functions of genetic algorithms with the maximum Lyapunov exponent,and optimized the weights of a neural network with the genetic algorithm.According to the fitness functions and the weights,a genetic algorithm neural network controller was designed,which improved the efficiency of neural network control.Numerical simulations were performed for the discrete system as Logistic mapping,continuous system as Rossler equations,and the chaotic transition of the AFM microcantilever.The results of numerical experiments indicated that the improved genetic neural network control method can control discrete and continuous chaos systems to the expected period orbit,and prove the reliability of the algorithm.