Abstract: In view of the increasingly serious social and ecological environmental problems caused by the rapid development of coalbed methane industry, this paper introduces the reachability matrix of ISM, triangular fuzzy number and nonlinear Hebbian learning algorithm into the fuzzy cognitive map model, and takes the coalbed methane exploitation area of Zhengzhuang block in Qinshui basin as an example.The results show that the traditional fuzzy cognitive map using expert method to obtain the correlation between indicators is subjective and difficult to obtain the correlation matrix has been improved, and the improved fuzzy cognitive map model is more accurate and stable.In the risk assessment grade of Zhengzhuang block, the maximum surface concentration of main atmospheric pollutants, the degree of production mechanization and the ratio of economic loss are the factors that have the greatest influence on the social ecological environment risk before, middle and later stage of coalbed methane development, respectively.