Smart grid multi-agent information has multi-source heterogeneity,the accuracy of effective use of information is low,and the fu-sion efficiency is poor.To solve the above problems,a smart grid multi-agent information automatic fusion model is constructed.The data warehouse technology is used to integrate the multi-agent information of smart grid and implement integrity processing and outlier processing.In order to reduce the amount and dimension of information,the information distance entropy feature is ex-tracted and the matching coefficient is calculated.Then,with the matching coefficient as the input value,the smart grid multi-agent information of the same category is divided together by K-means clustering algorithm to achieve automatic informa-tion fusion.The experimental results show that the consistency of information among the same kind of fusion reaches the maxi-mum value,which is more than 0.8;The time cost reaches the minimum value,which indicates that the fusion accuracy of the model is higher and the fusion efficiency is better.