The　modeling　process　is　resource-intensive,　time-consuming,　and　expensive　due　to　the　large　number　and　variety　of　production　and　auxiliary　equipment　in　discrete　manufacturing　plants.　The　popular　twin-model　construction　method　based　on　a　single　type　of　equipment　cannot　meet　the　demand　for　rapid　construction　and　high-quality　model　mapping　in　large　discrete　manufacturing　plants.　This　paper　proposes　a　strategy　for　developing　a　digital　twin　polymorphic　model　(DTPM)　for　discrete　manufacturing　workshops　to　meet　complex　manufacturing　business　scenarios,　improve　the　richness　of　digital　twin　model　varieties,　construction　efficiency　and　intelligence,　and　form　an　efficient　equipment　model　construction　method.　The　complete　element　information　and　real-time　dynamic　data　of　the　minimal　component　unit　(Functional　Components,　FC)　of　DTPM　are　clustered　and　analyzed.　In　addition,　properties　of　the　functional　component　information　model　are　characterized　from　several　dimensions,　such　as　geometry,　physics,　and　behavior.　Furthermore,　the　FC　with　highly　concentrated　primary　components　is　established　based　on　the　object-oriented　derivation　inheritance　and　attribute　reuse　hybrid　drive　technique.　FC　drastically　reduces　the　duration　of　the　digital　twin　model　development　process　through　encapsulation,　inheritance,　polymorphism,　and　other　technologies.　On　this　premise,　the　DTPM　construction　method　that　deeply　de　couples　the　physical　structure　and　functional　methods　is　proposed.　DTPM　integrates　adaptive　information　interaction　technology　to　accomplish　intelligent　data　connection,　dynamic　data　updating,　and　digitally　accurate　mapping　of　static-dynamic-virtual　multidimensional　models　in　discrete　production　workshops.　Lastly,　the　method＇s　validity　is　confirmed　by　creating　a　large　discrete　production　workplace.　The　results　indicate　that　the　proposed　technique　may　significantly　enhance　the　model　variety　and　building　efficiency　of　large-scale　digital　twin　workshop　systems.