The extensive application of modern information and communication technology in the power system through the in-depth integration of the information system and the power system has led to the gradual development of the cyber-physical power system (CPPS). While advanced information technology increases the safety and reliability of power system operations

it also increases the risks of fault propagation. To improve the reliability of CPPS from the perspective of power communication routing

it is proposed that the CPPS model and vulnerability assessment of power node reflect the correlation between information and energy flows with the service impact on power grid operation

which is an important index for evaluating communication services. According to the distribution of services at the different important levels on the links

the importance of the cross-layer link is established as the vulnerability evaluation index of the communication network. Then

the routing optimization model is proposed in combination with the service transmission risk under cyber-attack and the operating characteristics of the information system

which is solved through an improved fast-convergent genetic algorithm. The simulation results show that the proposed method allocates the alternate route to the low-risk link without significantly increasing the delay of the main route

which effectively improves the power supply reliability of CPPS in extreme cyber-attack scenarios.