As China 's coal mines enter deep mining, gas and water compound disasters have become the bottleneck of outburst safety. In the prevention and control project of high confined water in the floor of high outburst mine, cross-layer drilling is easy to induce the nonlinear coupling disaster of abnormal gas emission and high pressure water inrush. Based on the grouting reconstruction project of limestone aquifer in the upper section of Taiyuan Formation in No.22 coal seam floor of Chensilou Coal Mine, this paper analyzes the multi-physical field coupling disaster mechanism of gas and high-pressure water in the process of cross-layer drilling under the extreme conditions of mining elevation of-725-818 m, floor hydrostatic pressure of 1.79～ 2.72 MPa and normal fault of 5～ 12 m. On this basis, an intelligent collaborative early warning system architecture for gas-water composite disasters based on multi-source Internet of Things data fusion is constructed. Combined with the drilling engineering quantity of nearly 9 000 m and the final grouting pressure of 13 MPa in the mine, the field deployment scheme of the system was customized, and the high simulation digital twin scenario deduction was carried out for the exploration and injection holes exposing faults. The results show that the scheme can effectively fit the weak precursor of compound disaster caused by tectonic belts such as faults, and realize advanced and accurate four-level dynamic early warning and disaster prevention linkage. This study provides a forward-looking theoretical guidance and a landing technical blueprint for the compound treatment of floor water disaster and gas in deep high outburst mines, which can significantly improve the safety guarantee level of deep cross-layer drilling engineering.