Steel　structure　buildings　are　widely　favored　for　their　environmental　friendliness　and　exceptional　performance.　However,　traditional　methods　of　quality　risk　factor　assessment　are　limited　by　subjectivity　and　inefficiency.　To　address　this,　our　study　introduces　a　natural　language　processing　(NLP)　model　algorithm　to　identify　a　list　of　quality　risk　factors.　Initially,　quality　acceptance　and　accident　reports　of　403　prefabricated　steel　structure　buildings　were　processed　and　preprocessed.　Using　NLP　algorithms,　texts　were　successfully　clustered　into　themes,　yielding　five　thematic　results,　each　containing　ten　effective　keywords.　Through　in-depth　analysis　of　these　themes,　labels　for　each　theme　were　identified,　and　a　list　of　quality　risk　factors　was　compiled.　This　research　not　only　provides　a　new　method　of　indexing　quality　risk　for　steel　structures　but　also　significantly　enhances　the　sector＇s　digitization　and　intelligence.　This　advancement　is　crucial　for　the　development　of　the　steel　structure　building　industry,　aiding　in　more　efficient　and　accurate　identification　and　management　of　potential　quality　risks.