Abstract: For the traditional fault diagnosis method of power transformers cannot effectively obtain complete fault information, and thesystem model must be re-built when new constraints were added,a new fault diagnosis method of power transformer based on rough set theory and ASP was proposed.The gas dissolved in transformer oil is analyzed by rough set theory, then the problems to be solved are converted into ASP knowledge base by combining with ASP rules, and fault diagnosis are realized by the ASP solver. Compared with traditional systems, the diagnostic technique is simple and accurate, its expression ability is strong and has a certain flexibility and fault-tolerant ability, the dynamic maintenance of the knowledge base can be realized, and the accuracy rate of transformer fault diagnosis can be up to 94.8%.