Abstract: With the promotion and application of grid-connection of distributed energy, the DC microgrid as a carrier has many advantages in terms of controllability and power quality. Firstly, this paper aims at the problem of reliable communication control and power coordinated control in DC microgrid with multi-bus distributed energy. The uncertainty of photovoltaic and generator output is considered. The coefficient of power evaluation is defined. The rationalization of power coordination control is improved and the burden of multi-bus communication is reduced. Secondly, the long short-term memory network in deep learning is used to predict the photovoltaic output. Based on the distributed Raft algorithm to participate in the secondary control of the microgrid, the power coordinated control and economic optimization of the DC microgrid are realized. Finally, a simulation model of four-terminal DC microgrid based on PSCAD/EMTDC is established for validation. The simulation results show that the proposed control method can effectively improve microgrid stability and coordinate microgrid power control.