Rock burst is an important factor in the safety of tunnels and underground mines. In this phenomenon due to the concentration of stress and sudden explosion, the mass of rock is broken into small pieces and is thrown around, which usually causes distraction and great damage. with increasing depth of excavation and development of tunnels and underground mines, the potential of rock burst is the most important factor in the sustainability threat in these projects. In recent years the use of data mining methods has been extended to predict rock burst. In this research, the actual data of the rock burst related to different mines and tunnels worldwide have been used to study and evaluate this phenomenon. this data has been used to develop the support vector machine (SVM), k-nearest neighboring (KNN), bayesian networks (), artificial neural network (ANN) and the chaid tree model in IBM modeler 18 software. after the construction of models, the efficiency and precision of the prediction of data mining techniques were evaluated together with the three experimental methods of stress coefficient, brittleness coefficient and elastic energy index. The results show an optimal evaluation of data mining algorithms compared to empirical methods. Among the data mining algorithms, support vector machine, bayesian networks, and artificial neural networks have been present with the highest accuracy, respectively. the result of the analysis showed that considering the desirable evaluation of data mining models, can be used to evaluate the rock burst phenomenon.