The correct topological relationship in the low-voltage distribution network is crucial. The low-voltage distribution network structure changes frequently and tremendously because of the need of operation and maintenance. It is necessary to identify the topology of the incorrect updating of data, low circulation and poor quality, which can not correctly reflect the actual topological structure of low-voltage distribution network. Knowledge graph technology can clearly reflect the existing relationship between data, reason mining hidden knowledge, and apply to low-voltage distribution network topology identification. Therefore, this paper applies knowledge graph technology in topology identification. Firstly, it analyses the construction method of knowledge graph, uses knowledge graph technology to integrate data in multiple low-voltage distribution network information systems, deduces missing data, excavates the relationship between data, and then constructs the knowledge graph of low-voltage distribution network topology structure. Finally, combined with "Typical Design Specification of Low Voltage Distribution Network Infrastructure Project" and semantic word segmentation technology, the household transformer relationship in low voltage distribution network information system is verified and identified. The experimental results of the example are great, which show that the identification method proposed in this paper has theoretical value and practical application value. This work is supported by Special Project of China Southern Power Grid (No. ZBKJXM20170078) and National Natural Science Foundation of China (No. 51277134) “Mechanism and Method Research for Mass Electricity Physical Quantity Light-weight Transmission in Smart Grid”.