Abstract: In order to comprehensively improve the accuracy of monitoring and early warning of rock bursts in coal mines and improve the automation level of daily monitoring, taking the load conditions of rock bursts as the starting point, the idea of carrying out separate source monitoring for the load source of rock burst was proposed, and a comprehensive monitoring and early warning platform for the weight of separate source of rock burst in coal mine was developed. The results show that, based on the field monitoring data, attribute weights and grade weights were assigned to each early warning index, and then the comprehensive weights that can dynamically change with the impact risks were obtained. The comprehensive early warning method of rock burst source weight can solve the problem of subjective error caused by artificial weighting, avoid the neutralization of some indexes in the fixed weight, and the early warning results were more objective and reliable. A comprehensive early warning platform for rock burst has been developed, which integrated the functions of interface fusion, format conversion, statistical analysis, index priority, weight calculation and grade early warning. The comprehensive early warning platform for rock bursts has realized the in-depth development and intelligent release of multi parameter and multi-scale early warning information such as microseisms, ground noise, stress and drilling cuttings, etc. and the intelligent release of warning information, which greatly improved the efficiency and effect of early warning. The results have been applied in the bump-prone mines.