Abstract: The result of blast casting affects the stripping cost of open-pit mine directly，and has an important influence on the production efficiency to the stripping technology system of blast casting-dragline. In order to improve the accuracy of blast casting result prediction and feedback the design of parameters optimally. A prediction model(IPSO-ELM)was established for blast casting result in open-cast mine by combining of parameter-Improved Particle Swarm Optimization(IPSO)and Extreme Learning Machine(ELM)based on the analysis of the blast casting result influence factors. By adopting the parameter optimization method of Sigmoid inertial weight adaptive adjustment and learning factor dynamic adjustment，the shortcomings of low iteration efficiency and easy to fall into local convergence in the traditional particle swarm optimization are improved，and the model is analyzed by using the blasting data of the open-cast mine in Heidaigou，the prediction model of blast casting result was established by selecting powder factor，burden，bench height，hole spacing，minimum resistance line as the input variables，as well as effective stripping rate and loosen coefficient as the output variable，besides，the ELM model and the GA-ELM model were compared with it. The results show that the convergence speed of the model is enhanced obviously by improved adaptively and dynamically the inertia weight and learning factor of the conventional particle swarm optimization algorithm. IPSO-LMELM model is the feature of shorter running time and higher convergence speed，its average prediction accuracy reached 0.989 5，and the mean square error，mean absolute error and mean relative error were controlled at 0.271 4，5.056 5% and 0.260 3%. The prediction error of IPSO-ELM is significantly lower than ELM and GA-ELM prediction model and the prediction accuracy is the highest.IPSO-ELM has a better applicability for predicting the effect of blast casting in open-pit mines.