Abstract: Existing problems on path planning based on fuzzy logic for mobile robots are analyzed. The kinematic model of a mobile robot is built up, and a self organized fuzzy controller applied to path planning in unknown, unstructured environment is designed. Establishment of fuzzy rules accords well with the key idea of greedy algorithm. Some new methods are presented to solve such issues like "symmetric indecision", deadlock estimation, freedom from deadlock state and the judgement whether deadlock is eliminated or not. Simulations in several kinds of environments verify the validity and reasonability of these methods.