Abstract: In order to implement more effective self-monitoring and self-diagnosis for smart distribution network, a health level evaluation method of smart distribution network based on multivariate information was proposed. Firstly, combined with the smart distribution network information system, the health evaluation index system of equipments and grid structure related to the health level of smart distribution network was constructed. Secondly, according to the multivariate index data, the improved neural network was used to calculate the health index of equipments and grid, and the health index of equipment layer was solved based on grid topology and goal oriented method. Then, according to the health information of the equipment layer and the grid, combined with fuzzy theory and D−S evidence theory, fuzzy evaluation and uncertain reasoning were carried out to obtain the membership degree of each health level of the distribution network, and the overall health level of the smart distribution network was evaluated. The example results show that the proposed evaluation method can make a rational evaluation of the health level of the distribution network by integrating the multi-information such as the equipment index, grid index and the grid topology structure of the smart distribution network, realize the self-monitoring and self-diagnosis of the smart distribution network, and accurately reflect the impact of equipment maintenance and replacement and grid optimization on the health level of the distribution network.