Abstract: When conventional genetic algorithm(GA) is used to cope with some complex problems, slow convergence or prematurity often occurs. A novel evolutionary algorithm, based on the rational decision-making of human, the rational genetic algorithm (RGA) is proposed to solve these problems. The key point of RGA is to use the genetic information feedback and set up rational rules to guide the evolution of genetic individuals. The proposed RGA effectively incorporates inheriting and learning behaviors of knowledge and experiences of species into GA. The problem of multi-robot motion cooperation under known circumstance can be solved better by RGA than conventional GA. Theoretical analysis and simulation results show the validity of RGA.