With　the　rapid　development　of　Internet　of　Things　(IoT),　sensors　play　an　indispensable　role　in　the　field　of　monitoring.　In　monitoring,　the　data　collected　by　sensors　as　a　part　of　the　monitoring　and　control　systems,　and　the　data　quality　directly　impacts　the　accuracy　of　the　analysis.　However,　some　factors　such　as　environmental　changes,　node　failures　and　structure　damage　can　result　in　sensor　abnormalities　like　frequent　topology　changes,　communication　failures,　which　greatly　reduces　sensor　data　reliability.　To　overcome　this　challenge,　we　propose　to　utilize　the　mathematical　models　to　adaptively　correct　the　deviated　sensors　data.　Firstly,　an　initial　evaluation　of　abnormal　sensor　is　made　based　on　the　changing　trends.　Then,　mathematical　models　are　formulated　for　each　sensor　based　on　the　monitoring　data,　and　these　models　are　employed　to　rectify　the　deviations　in　the　monitoring　data.　Moreover,　the　corrected　data　is　validated　based　on　the　observed　trends.　Experiments　are　executed　based　on　sensors　data　collected　from　a　real　application　scenario.　The　results　confirm　that　the　proposed　correction　method　can　significantly　correct　the　sensors　deviation　data　of　monitoring.　©　2023　IEEE.