Abstract: To improve the accuracy and reliability of rockburst level prediction, an asymmetric variable-weight cloud model for rockburst prediction is presented. Considering the differences in the connotation and interrelation of the prediction indexes of the rockburst intensity level, the system of rockburst intensity level prediction indexes including four indexes such as brittleness coefficient, stress coefficient, elastic deformation energy index and integrity coefficient of the rock body is established by adopting the method of literature research and the method of frequency statistics. Fully utilizing the differences in attribute feature information of rockburst objects and the recognition and judgment information of experts on the complexity and variability of engineering conditions, the asymmetric variable-weight cloud model for rockburst prediction is developed to reduce the subjective arbitrariness in the calculation of rockburst prediction weights, reflect objectively the impact of different attribute characteristics of rock burst on variable-weight calculation, describe effectively the random uncertainty and fuzzy uncertainty of small changes at the boundary of rock burst prediction levels, as well as the characteristic information of single boundaries in their edge intervals, and also improves the reliability and precision of the prediction results of the rock burst intensity levels by using fuzzy hierarchical analysis, anti-entropy weighting method and cloud modeling method based on the ideas of game theory and variable-weight. The model has been applied to 20 sets of domestic and international rockburst examples. The prediction accuracy is 90% and higher than the prediction accuracy of entropy weight-cloud model, improved CRITIC-multidimensional cloud model and RS-TOPSIS method, which verified the reliability and accuracy of the model. The establishment of the model in this paper provides a method path for objectively reflecting the actual situation of rockburst engineering and improving the reliability and accuracy of rockburst intensity level prediction results.