SF6 gas is widely used in various electrical equipment due to its excellent insulation and arc-extinguishing properties. Research on the monitoring and early warning of SF6 leaks in electrical equipment is crucial for the safe and stable operation of power systems and a significant step towards building ecological civilization and achieving carbon neutrality. To address the issue of direct carbon emissions from SF6 leaks in practical engineering, this paper presents a method for accounting for direct carbon emissions from SF6 based on spatiotemporal analysis of pressure values. This method enables efficient carbon emission accounting for SF6 leaks. A three-dimensional discrimination model for high car-bon emission alerts has been developed to accurately identify devices with high emissions. Simulation and experimental studies on direct carbon emissions from SF6 have been conducted, with simulation predictions and experimental errors within 20%, indicating a need for further optimization. This research can effectively contribute to the construction of green, low-carbon power grids.