With the development of Energy Internet in recent years, the rapid promotion of new energy emerging facilities highlights the trend of the future energy interconnection. Under the system architecture, renewable energy and grid, energy storage, energy deployment and transportation problems require a flexible grid scheduling strategy to deal with complex consumer tasks. The optimization of electric-gas-thermal mixed energy supply system is an effective way to solve the rational allocation of energy. Optimizing the network power flow can not only improve the efficiency of energy utilization, Optimizing the network power flow can not only improve the efficiency of energy use, but also minimize network redundancy through the high degree of coupling. In this paper, the mathematical model of multi-energy load is established under the framework of advanced measurement system and energy Internet. Based on the load forecasting of LMBP algorithm, an optimal genetic algorithm is used to solve the establishment of an example to achieve optimal allocation of multi-energy flow and capacity allocation.