Accurate and dynamic accounting of power carbon emission factors (PCEFs) is essential for supporting low-carbon transition and ensuring the integrity of green power markets. Traditional static and province-level approaches compress spatiotemporal variations into annual averages, which obscures renewable energy volatility, cross-regional power flows, and the environmental attributes of traded green power. This study develops a high-resolution, time- and zone-specific PCEF model that integrates unit-level generation data, hourly granularity, interregional transmission, and a green power deduction mechanism to prevent double counting. Using the North Hebei power grid as the primary case, characterized by over 80% renewable capacity and large-scale clean power exports, the model demonstrates significant improvements in capturing intra-day and seasonal dynamics of carbon intensity. Results show that the mixed PCEF with hybrid power and regional exchange in North Hebei (0.5069 tCO₂/MWh) is notably lower than the fossil-fuel baseline (0.7899 tCO₂/MWh), while the deduction of green power trading raises the retained local factor to 0.6488 tCO₂/MWh. Comparative analysis with Jiangsu Province, a region dominated by fossil power but with high external clean power inflows, validates the model’s robustness across diverse energy structures. The findings highlight three key contributions: (1) improving spatiotemporal resolution of PCEF calculations, (2) clarifying carbon responsibility allocation in cross-regional flows, and (3) enhancing the credibility of green power trading mechanisms. This research provides methodological and empirical evidence to guide the development of unified carbon accounting standards, optimize power trading, and support policy design for equitable and effective decarbonization in China’s power sector and beyond.