A　patent＇s　prior　knowledge　is　its　related　pre-existing　technological　knowledge,　i.e.,　its　search　space.　Accurately　organizing　and　describing　the　prior　knowledge　of　patents　is　essential　for　gaining　in-depth　understanding　of　the　sources　of　invention　knowledge.　In　this　paper,　we　propose　a　method　to　construct　a　prior　knowledge　network　using　the　abstract　text　of　relevant　patents　before　the　focal　patent　disclosure.　We　then　analyze　the　structure　of　the　prior　knowledge　network　based　on　a　sample　of　410　patents　in　the　field　of　lithography　technology.　Our　research　shows　that　the　prior　knowledge　network　is　a　fast　network　of　information　transfer　and　flow,　exhibiting　scale-free　characteristics　and　small-world　characteristics.　Furthermore,　by　comparing　the　network　structure　characteristics　of　high-value　and　low-value　patents,　we　find　that　high-value　patents　have　a　larger　network　size,　higher　degree　centrality　of　problem　knowledge　elements　(PKEs)　and　solution　knowledge　elements　(SKEs),　as　well　as　smaller　network　density,　average　shortest　path　length　and　average　shortest　path　length　between　PKEs　and　SKEs.　However,　these　differences　are　not　statistically　significant.　©　2023　ACM.