This paper described a knowledge based system that can aid in the diagnosis of sealed coal mine fires. Using a commercially available shell, the expert system has been designed to accommodate a series of diagnostic rules replicating the reasoning logic that would be used by human experts. The system allows the diagnosis of explosibility of the fire zone and trends of various fire ratios to be easily visualized for decision making. By continuously analyzing the incoming information, the system is able to provide a rapid assessment of the situation of a sealed mine fire prevailing under specific mining and geological conditions. In addition, short term trends in fire behavior can be predicted and possible fire control actions recommended. An example is given and shows how this system can be used as an analytical tool in various ways to assist the combating of coal mine fires. Further development of the system is discussed.