Abstract: The gas of adjacent layers is migrated to the fully mechanized working face by slit channel.So,it is one of the important elements of transfinite gas in mine.The BP neural network is improved by the genetic algorithm,so as to obtain the prediction model of adjacent layers of gas emisson.According to the degree importance of structure,the grey correlation is adopted to make the prediction results of fully mechanized working face more exact.So,the research results can contribute to the gas controlling in the fully mechanized working face. 　　

Key words: Grey correlation theory, Neural network, Adjacent layers, Gas emission