After a coal mine disaster, especially a gas and coal dust explosion, the space-restricted and unstructured underground terrain and explosive gas require coal mine rescue robots with good obstacle-surmounting performance. This paper tackles path-planning or collision avoidance problem. The collision avoidance problem in path planning of mobile robots was inspired from fuzzy control concept. The deduction of the path-planning algorithm is followed. Fuzzy rules and fuzzy inference based on experiences is built, it constructed a reasonable and applicable control reactive rule table. The simulation results are presented to validate the approach. And it is also proved that the control arithmetic is available in both static and dynamic obstacles in coal mine environment.