Maintaining　the　effective　operation　of　chillers　depends　heavily　on　prompt　and　precise　fault　diagnosis.　A　chiller　fault　diagnosis　method　based　on　attention　and　long　and　short-term　memory　fully　convolutional　neural　network　(ALSTM-FCN)　was　proposed　in　light　of　the　nonlinearity　of　chillers＇　operation　data　and　the　challenge　of　fault　diagnosis　models　in　identifying　effective　features,　which　results　in　low　model　accuracy.　First,　an　FCN　fault　diagnosis　model　was　established　according　to　the　feature　selection　results,　and　the　proposed　Lastly,　the　attention　mechanism　was　included　to　the　model　with　the　intention　of　emphasizing　the　significance　of　features　and　enhancing　the　accuracy　of　chiller　problem　diagnosis.　Furthermore,　the　ASHRAE　1043-RP　dataset　was　used　to　confirm　the　efficacy　of　the　suggested　approach.　©　2024　IEEE.