Real-time　changes　in　the　mechanical　parameters　of　prestressed　steel　structures　can　limit　the　accuracy　of　the　analysis　and　prediction　of　their　performance，which　in　turn　presents　significant　challenges　to　developing　intelligent　safety　control　systems　for　these　structures.　This　study　proposes　a　digital　twin-based　intelligent　safety　control　system　for　prestressed　steel　structures.　An　intelligent　safety　control　framework　based　on　digital　twins　is　established，considering　the　spatiotemporal　evolution　of　structural　safety.　The　safety　control　system　is　divided　into　two　aspects：performance　analysis　and　prediction.　The　construction　of　the　twin　model　is　analyzed，and　the　key　node　coordinates　are　corrected　based　on　the　control　framework，three-dimensional　laser　scanning，and　the　weighted　average　method　to　reduce　errors　in　the　simulation.　The　weighted　average　method　is　used　to　comprehensively　analyze　the　force　on　each　node　of　the　cable，and　the　component　size　in　the　correction　model　can　effectively　reduce　the　simulation　errors　caused　by　material　defects.　Accurate　mechanical　parameters　of　structural　safety　can　be　obtained　using　the　twin　model，providing　data　support　for　the　analysis　and　prediction　of　structural　safety.　The　D-S　evidence　theory　is　integrated　to　analyze　the　degree　of　safety　of　the　structure　as　well　as　to　determine　the　most　critical　force　components　and　nodes　with　maximum　change　in　their　mechanical　parameters.　A　method　that　combines　digital　twins　and　random　forest　techniques　is　established　to　analyze　the　influence　of　various　factors　on　structural　safety.　The　safety　of　the　structure　is　predicted　by　adjusting　the　key　influencing　factors，and　the　structure　safety　control　mechanism　is　finally　formed.　Based　on　this，the　safety　of　the　structure　can　be　ensured　by　implementing　structural　safety　maintenance　measures.　The　effectiveness　of　digital　twins　in　structural　safety　control　was　verified　by　applying　it　to　the　spoke-type　cable　truss.　©　2023　Tianjin　University.　All　rights　reserved.