Abstract:Concentrating on input-output method, we propose the computing methods of direct carbon emission coefficient, complete carbon emission coefficient and embodied carbon of export trade. With the manufacturing industry of China as an example, we have measured the direct carbon emission coefficient, complete carbon emission coefficient and embodied carbon of export trade, and used structure decomposition analysis to disintegrate manufacturing export trade implicit carbon variation into: technical effect, structure effect and scale effect. we achieved a series of results: (1) Direct carbon emission coefficient and complete carbon emission coefficient all show a trend of decline, it shows that the policy of energy conservation and emissions reduction has achieved good results; (2) Export embodied carbon percentage points more than 50% of total carbon dioxide emissions between 2000-2007, it shows that more than half the carbon dioxide emissions are caused by the import demand of China's trading partners;(3) Embodied carbon of export trade of manufacturing industry shows a trend of increase. To reduce domestic carbon emissions, we should improve the production technology of main export department, and select appropriate alternatives for production and export; (4)Scale effect has a positive effect on the increase of export trade implicit carbon emissions, but the technology effect has a negative effect on it. At last, we give the suggestions.