The purpose of this paper is to study the current high implementation cost of distributed transaction scheduling in the micro-grid, and the difference in transaction demand caused by the impact of the external environment on the amount and output of clean energy in different transaction scenarios, resulting in the problem that new energy cannot be fully absorbed. A smart transaction scheduling model of micro-grid considering carbon emission certification is proposed in this paper. Firstly, carbon performance functions and multi-time scale combined consensus functions are designed to reduce transaction costs and reduce computing power consumption. Secondly, the scheduling plan is formulated under the weak centralized scheduling architecture based on the public network of the micro-grid, and an evolutionary game algorithm with a joint incentive mechanism is designed to solve it, improving the capacity of new energy consumption. Finally, the data of Beijing Jin-feng Energy Internet Park are utilized to analyze the calculation example, and the correlation laws between reliability, decentralization, micro-grid cost and efficiency are obtained.