Abstract: Aiming at the sequential collaborative optimization problem of the electric power and traffic network integration network (IETN) with dynamic interaction of power distribution system, transportation system and electric vehicle under multiple uncertain factors, a collaborative optimization method for IETN based on graph neural network multi-agent reinforcement learning is proposed to solve the above problems. Firstly, the interaction relationship among electric vehicles is converted into a dynamic network graph model based on the graph theory. A graph neural network multi-agent reinforcement learning algorithm based on attention mechanism is proposed to solve the charging navigation strategy of electric vehicles, so as to explore the mutual influence among electric vehicles. Then, the optimal power flow of distribution network considering the renewable energy output in an active distribution network is solved, and the marginal price of distribution network node is obtained by the second-order cone optimization and dual optimization theory, so as to study the dynamic interaction characteristics of power and transportation systems. Finally, the simulation verification is carried out on a 108-node traffic network and the IEEE 33-node power system in a certain area.