This paper introduces the optimization methods of Genetic Algorithm.Based on the Different Location Crossover and the Same Location Crossover,a new leading crossover is proposed.Then it is a self-adaptive manner judgment to choose which crossover is used before the crossover operator.At last,five different tests of the simulation function are given.The results show that the leading crossover is more efficient to improve convergence than other crossovers.And the new method is easy to find the optimal solution.