Abstract: Gas disasters are still the main disasters during deep coal mining in China. With the increase of mining depth and mining intensity, the gas disasters are becoming more serious and happen occasionally, owing to the complex coupling relationship between the inducing factors of gas disasters. Therefore, integrated analysis and early warning of potential risks of gas disasters, based on monitoring big data, have become urgent demands for coal mine safety management and supervision. In this paper, the big data characteristics of the potential risks for gas disasters were analyzed and the early warning methods were put forward. Additionally, the big data monitoring and early warning cloud platform for gas disasters, which can display the monitoring and early warning information in “one picture” timely, was developed and applied in coalmines. Besides, through the in-depth analysis of all relevant real-time data, variation tendency and their relationship, the platform can also realize that the big data monitoring, automatic identification, early warning and SMS sending of results for gas disasters.