Abstract: In order to simplify the prediction process of mine pre-split blasting effect and improve the accuracy of prediction，in view of the traditional effect of pre-split blasting，which focuses on the defects of pre-split slope formation，considering that the slope of open-pit mine is often affected by the dynamic load such as blasting rock vibration，a neural network prediction method of mine pre-splitting effect based on considering both the slope forming standard and the impact of blasting vibration on slope is proposed.The parameters such as single hole charge，average hole depth，hole spacing，peak vibration velocity （horizontal，vertical），main vibration frequency （horizontal，vertical），blasting distance and so on are used as input parameters of neural network.Parameters such as average vibration speed，half-hole rate，roughness and fracture coefficient are taken as the output parameters of the neural network.An prediction method of mine pre-splitting blasting effect is established based on 24 blasting technical data of adjacent slope.3 on-site blasting prediction test results show that the rock properties，engineering geological conditions an other factors related to the control of pre-splitting blasting effect are simplified and integrated into the neural network by its self-organizing structure.Therefore，the relative error of the average vibration velocity can be controlled at about 7% by the trained neural network using this data type，and the relative error of the half-porosity，roughness and fracture coefficient can be controlled at about 3%.

Key words: Open-pit mining, Pre-splitting blasting, Effect prediction, BP neural network, Prediction accuracy