Abstract: In the research and development of intelligent operation ticket system, the design of power transfer operation ticket needs to evaluate the power outage scope and line transfer path of distribution network failure or maintenance. Considering the actual factors of super large scale of power grid, the conventional relational database is difficult to support the real-time division of power grid and the decision-making of power transfer scheme. In order to solve this problem, by combining Dijkstra, deep/breadth first search and other graph theory algorithms, a decision-making scheme for the division of large-scale power grid fault affected areas and switching power supply based on graph database is proposed, which realizes the optimal path planning in the process of power dispatching, and is applied to the design of power dispatching operation ticket to meet the design requirements of fast and accurate invoicing. In addition to being applied in the operation ticket system, this scheme can also be applied to other occasions. For example, when the power grid needs power regulation due to faults or maintenance, the optimal power regulation scheme can be directly given for the use of regulators, which can save the economic cost of power grid dispatching, reduce the power loss in the process of power transmission, and avoid the wrong decision-making of power dispatching caused by human factors, which is consistent with the carbon neutrality goal of China’s power industry at present. It provides a feasible reference case for the automatic planning and selection of ultra-large-scale power grid transfer scheme.