This paper employed the multi-regional input-output (MRIO) to clarify the spatial and temporal evolution of the embodied pollutant emission patterns in China's interprovincial trade in 2012, 2015, and 2017. The structural decomposition analysis (SDA) model was applied to identify the socio-economic factors affecting the changes in pollutant emissions. The results showed that the areas with net pollutant export gradually shifted from those which have developed economic on the east coast to those have rich resources in the northwest. Moreover, the gap between provinces in net imports and exports of pollutants decreased. For example, the gap in SO₂ from 2012 to 2017 reduced by 63%. In addition, the increase of final demand was the main driver of the increase of emissions, with an average contribution of up to 669% to the emissions of the four air pollutants. The optimization of the energy consumption structure and the reduction of energy consumption per unit of GDP effectively curbed the momentum of emissions increase. The average contribution of the energy consumption structure to the four pollutants decreased from -220.75% to -546.25%. Although the impact of production technologies is minor, its impact on emissions has shifted from a facilitating to a dampening effect. This study provides new insights for improving energy efficiency and developing a clean energy mix, which contributes to developing measures to mitigate demand-side pollutant emissions and coordinated regional trade policies.