Monitoring　systems　for　the　safety　of　building　structure　(SBS)　can　provide　people　with　important　data　related　to　main　supporting　points　in　a　building　and　then　help　people　to　make　a　reasonable　maintenance　schedule.　However,　more　and　more　data　bring　a　challenge　for　data　management　and　data　mining.　In　order　to　meet　this　challenge,　under　the　framework　of　Wisdom　Web　of　Things　(W2T),　we　design　a　monitoring　system　for　the　SBS　by　using　the　semantic　and　the　multisource　data　fusion　technologies.　This　system　establishes　a　dynamical　data　cycle　among　the　physical　world　(buildings),　the　social　world　(humans),　and　the　cyber　world　(computers)　and　provides　various　services　in　the　monitoring　process　to　alleviate　engineers＇　workload.　Furthermore,　all　data　in　the　cyber　world　are　organized　as　the　raw　data,　the　semantic　information,　and　the　multisource　knowledge.　Based　on　this　organization,　we　can　concentrate　on　the　data　fusion　from　the　viewpoints　of　time,　space,　and　multisensor.　At　last,　a　prototype　system　powered　by　the　semantic　platform　LarKC　is　tested　from　the　aspects　of　sample　performance　and　time　consumption.　In　particular,　noisy　data　(i.e.,　inconsistent,　abnormal,　or　error　data)　are　detected　through　the　fusion　of　multisource　knowledge,　and　some　rule-based　reasoning　is　conducted　to　provide　personalized　service.