Abstract: At present, the research on partial discharge of transformer usually adopts a single partial discharge detection method to monitor and evaluate the condition of transformer. For the transformer with relatively complex structure, such detection technology and algorithm have false detection and missed detection, resulting in unreasonable transformer condition evaluation and poor engineering application. Therefore, a transformer condition monitoring and evaluation strategy based on partial discharge information fusion is proposed in this paper. For the same discharge power supply, the transformer operation state is evaluated through the information fusion of three partial discharge detection methods: ultrasonic, UHF and high-frequency current. At the same time, in order to overcome the decisive influence of subjective or objective factors on the evaluation results, this paper uses the combined weighting method of subjective weighting and objective weighting to obtain the comprehensive weight, and uses the matter-element analysis method in extension to establish the transformer state evaluation model and correlation matrix to determine the evaluation level of transformer operation state. The example calculation shows that the method based on partial discharge information fusion can reduce the uncertainty and subjectivity in the evaluation process, so as to obtain more reasonable and objective evaluation results.