Conventional genetic algorithms use only one crossover operator to generate the next generation. Different crossover operators, however, are suitable for different problems, even for different stages of the genetic process in a problem. Determining which crossover operators should be used is quite difficult and is usually done by trial-and-error. In this paper, a new genetic algorithm, the dynamic crossover genetic algorithm (DCGA), is proposed to solve the problem. The dynamic crossover genetic algorithm uses more than one crossover operator to generate the next generation. The crossover ratio of each operator changes along with the evaluation results from the respective offspring in the next generation. Several methods are proposed to adjust the control rate of each crossover operator. We thus expect that very good operators will have an increased effect on the genetic process. Experiments were also conducted, with results showing that the proposed algorithm always performs better than most (even all) algorithms having a single crossover operator.