Abstract: The deep coupling and complementary coordination of heterogeneous energy flows such as electricity and natural gas hare made the multi-area interconnected electricity-gas coupling system for renewable energy consumption become an essential direction for the development of Energy Internet in the future. In this context, comprehensively considering the multiple uncertain factors such as wind power output, gas consumption of gas-fired units and the interaction energy flow through tie-lines between distribution networks, a decentralized coordinated optimal scheduling framework based on the multi-agent autonomous decision for the multi-area integrated energy system is proposed. Firstly, a distributionally robust optimal scheduling model based on the ambiguity set of wind power output is established for the distribution network. Combined with duality theory and linear decision rules, the above semi-infinite optimization problem is transformed into a linear programming problem. Secondly, the second-order cone programming model for gas network is built, and according to the uncertain set of gas consumption for gas-fired units, the feasibility of its reserve configuration in the distribution network is tested. Then, the decoupling mechanism of energy flows in multi-area interconnected system is analyzed and an alternative direction method of multipliers based on parallel regularization is developed to solve the decentralized coordinated scheduling model for multi-area electricity-gas coupling system. Finally, the case results demonstrate the effectiveness and practicability of the proposed model and solution method.