<p>The digital and intelligent transformation for safety emergency represents a pivotal strategy for coal mining enterprises to enhance emergency management capabilities and mitigate accident risks. However, this transformation is often hindered by substantial initial costs and long investment return cycles. Government subsidies, as a key policy tool, play a crucial role in guiding and incentivizing enterprises to pursue this transformation. From a safety oversight perspective, this study employs evolutionary game theory to construct a game model between coal mining enterprises and the mine safety supervisory authority. It systematically analyzes the intrinsic mechanisms of strategic choices and dynamic evolutionary patterns of players under three mechanisms: static subsidy, initial dynamic subsidy, and optimized dynamic subsidy. Furthermore, by integrating complex network theory, this study establishes a diffusion model to explore the diffusion mechanisms and spatial distribution characteristics of transformation behaviors across the industrial network. The research demonstrates that: (1) The dynamic nature of the subsidy mechanism is crucial for system stability. (2) Under the dynamic subsidy mechanism, the initial subsidy coefficient is negatively correlated with the transformation’s effectiveness. Excessive subsidies can trigger an incentive paradox, suggesting an optimal range for the initial subsidy coefficient. (3) The diffusion process of the digital and intelligent transformation for safety emergency in coal mining enterprises under dynamic subsidy mechanisms exhibits three-phase characteristics.</p>