In order to solve the function optimization problem, an adaptive genetic algorithm is proposed based on the structure of the chain agent. The algorithm adopts the structure of the chain agent to reduce the computational overhead. The mixing crossover strategy that includes the arithmetic crossover and two-point crossover is adopted in the crossover operator. In the crossover and mutation operator, the adaptive probabilities of crossover and mutation related to the generations are adopted. The simulations reveal that this algorithm has better convergence and the speed to get the optimal solution is very fast.