Abstract: For problem of poor forecast effect of existing gas forecast expert system of coal mine because of no new knowledge acquisition measures and self-renewal function for knowledge, a knowledge acquisition approach based on rough sets theory was proposed. The method firstly establishes forecast samples of relationship between gas data and gas outburst intensity; then uses algorithms of continuous attribute discretization, attribute reduction and rules extraction based on rough sets theory to obtain forecast knowledge from lots of forecast samples, and stores the knowledge in knowledge database of expert system; finally, realizes real-time gas forecast based on reasoning machine. Example analysis result verifies effectiveness and practicality of rough set method applied in knowledge acquisition of gas forecast expert system of coal mine.