For the problem that current coal mine safety monitoring systems cannot remind workers of abnormal gas changes when the gas concentration is at a low level caused by lacking of analysis function of gas trends, a gas abnormality early-warning system of coal mine was designed. Monitoring data of safety monitoring system and integrated automation system is accessed by approach of file and OPC. Combined with actual mine ventilation, reference value and state of early-warning of checked points are calculated by fixed threshold model, mean square deviation early-warning model and time-sharing mean square deviation early-warning model. And early-warning information is released through message client, short messages and other means. The application showed the system achieves effective prevention of gas disasters and provides a reference for ensuring safety of underground workers.