B-pillar　structural　part　is　the　key　component　of　lightweight　automobile　body　structure.　To　possess　good　strength　and　plastic　fit,　B-pillar　generally　has　gradient　characteristics　of　mechanical　properties　along　its　length,　and　there　is　a　characteristic　transitional　range　from　high　strength　and　low　elongation　hard　parts　to　low　strength　and　high　elongation　soft　parts.　To　detect　the　local　mechanical　properties,　the　destructive　sampling　method　is　conventionally　used　for　evaluating　the　mechanical　properties　of　the　transition　region　of　automobile　B-pillar.　It　is　time-consuming　and　impossible　to　directly　conduct　rapid　non-destructive　testing　of　parts.　Hence,　a　micromagnetic　automatic　detection　system　for　the　B-pillar　of　the　body　structure　is　developed　to　realize　the　non-destructive　evaluation　of　the　mechanical　properties　of　the　transition　region.　The　BP　neural　network　model　is　established　by　the　magnetic　characteristic　parameters　and　mechanical　performance　indicators　obtained　from　the　B-pillar.　It　helps　to　realize　the　quantitative　characterization　of　the　mechanical　properties　of　the　transition　region　of　the　B-pillar　structural　part　of　an　automobile,　which　has　important　engineering　application　prospects.　©　2024,　The　Author(s),　under　exclusive　license　to　Springer　Nature　Switzerland　AG.