Abstract: The effective stripping ratio is one of the most important indicators to evaluate the effect of casting blast，and the prediction of effective stripping ratio can guide the technical personnel to make accurate production planning.Based on analyzing the factors of effective stripping ratio，the bench height，explosive consumption，bottom burden，hole spacing，row spacing，thickness of coal seam were taken as the network input of generalized regression neural network (GRNN)，the effective stripping ratio was set as network output.The prediction model of effective stripping ratio was built，and network training and prediction were achieved through MATLAB programming.The results showed that the GRNN could predict the effective stripping ratio accurately with the forecast error normally around 5%，and the prediction results can meet the project requirements.

Key words: Casting blast, Effective stripping ratio, Reliability, Generalized regression neural network