At　present,　with　the　rapid　development　of　Internet　information　technology,　the　demand　for　enterprise-level　intelligent　information　technology　software　projects　has　emerged　in　large　numbers.　Software　trustworthiness　has　become　an　important　part　of　the　development　of　the　software　industry　that　cannot　be　ignored.　For　developers,　software　trustworthiness　should　be　considered　as　an　important　factor　from　software　development　to　the　end.　Software　trustworthiness　exists　in　the　whole　life　cycle　of　software,　at　present,　there　is　no　trustworthy　declaration　document　that　can　be　applied　in　the　whole　development　process　of　software,　which　makes　it　necessary　for　developers　to　repeatedly　define　trustworthy　declaration　documents　in　the　process　of　developing　information　technology　projects　with　the　same　functional　attributes.　Based　on　the　above　existing　problems,　this　study　takes　the　front　and　back　ends　of　web　and　WeChat　applets　as　the　research　objects.　By　obtaining　the　functional　requirement　documents　of　informatization　projects　as　the　data　base,　introducing　natural　language　processing　technology,　constructing　a　semantic　analysis　model　conforming　to　this　study　through　semantic　analysis　processing　technology,　and　finally　realizing　the　function　of　automatically　generating　trustworthy　behavior　declaration　files　by　traversing　the　language　under　the　graph　computing　framework.　In　this　thesis,　we　focus　on　the　functional　attribute　extraction　method　based　on　semantic　analysis　technology　and　the　knowledge　graph　generation　and　storage　method　of　functional　attribute　keywords　and　high-frequency　words,　which　lay　the　foundation　for　the　automatic　generation　of　trustworthy　behavior　statement　documents　afterwards.　©　2023　SPIE.