Abstract: With the help of support vector machine model and considering the factors that affect the distance of forward replacement and backward replacement，the prediction model of support vector machine（SVM）for the forward and backward displacement of blasted pile was proposed. Combined with the engineering examples and blasting operation patterns， this experiment selects the hole depth，hole distance，resistance line distance，slope angle，over drilled length，charge quantity，stemming length of the first row of blasthole，as well as the hole depth，hole distance，row distance，over drilled length，charge quantity and stemming length of the second to eighth rows of blastholes as the influencing factors，and takes the forward and backward replacement distance of the blasting pile as the output variables. In addition，GSM（grid search method）and GA（genetic algorithm）are used to optimize and adjust the hyper-parameters of SVM prediction models. 40 groups of blasting examples collected from the field blasting operation are used to develop and evaluate the prediction model. The final prediction results show that there is a certain mapping relationship between the relevant parameters collected from the field and the blasting pile replacement，and the support vector machine prediction model of blasting displacement combined with heuristic algorithm can get higher prediction accuracy.

Key words: Forward blasting pile, Backward blasting file, Support vector machine, Prediction