To enhance the accuracy of the fault diagnosis of power transformer and adapt to the development trend of power equipment condition, and overcome the limitations of traditional intelligent algorithms, a method of comprehensive evaluation of power transformer status based on association rules and variable weight coefficients is proposed in this paper. According to the experimental tests and monitored data, the evolution system is established. Through 9 comprehensive diagnosis faults and 24 single fault status, the belief confidence and variable weight coefficients calculated by the correlation rules, as well as the relationship between the running status of power transformer and the related maintenance records, the overall performance of power transformer can be derived. The experimental results reveal that the accuracy of the method can reach to 91%, which can accurately assess the health status of the power transformer. Also, the proposed method can guide the maintenance personnel to service equipment and provide reference for the status evaluation of the other high-voltage power equipment. This work is supported by National Key Research and Development Program of China (No. 2018YFB5100803), National Natural Science Foundation of China (No. 51606035), Project of Science and Technology Department of Jilin Province (No. 20190302079GX), and Project of Education Department of Yunnan Province (No. 2016ZDX261).