Abstract: An adaptive genetic algorithm for the optimum path planning problem of a mobile robot was proposed. The research project was carried out from four aspects:a geometry obstacle avoiding algorithm was developed to generate initial population; the crossover, mutation, improving and deletion operators which base on heuristic knowledge were designed for path planning; a new kind of fuzzy logic control algorithm was adopted to self-adaptively adjust the probabilities of crossover and mutation; simulation studies in both off-line and on-line environments were implemented. The simulation results show that the adaptive genetic algorithm has advantages such as rapid search speed, high search quality and strong self-adaptability. It is a new approach for solving the optimum path planning problem of a mobile robot.