The　interlayer　component　is　the　weak　layer　of　the　composite　beam.　For　the　post-strengthened　beam,　the　upper　and　lower　layers　are　connected　by　glue　bonding.　The　glue　is　mainly　applied　manually,　and　the　craft　has　great　randomness,　and　the　randomness　has　spatial　variability,　which　affect　the　normal　use　and　bearing　capacity　reliability　of　the　composite　beams.　In　order　to　study　the　probability　response　of　composite　beams　considering　the　spatial　random　field　characteristics　of　interlayer　materials,　a　finite　element　model　was　established.　The　upper　and　lower　beams　were　both　Euler　beam,　and　the　interlayer　components　were　regarded　as　continuous　springs,　which　were　discretized　and　used　as　springs　to　connect　the　upper　and　lower　beams　into　system　stiffness.　The　random　field　characteristics　of　the　continuous　spring　were　expressed　by　Karhunen-Loeve　expansion,　and　the　statistical　moments　of　the　response　were　calculated　by　the　reduced-dimension　point　estimation　method,　forming　the　calculation　strategy　of　KLE-PEM.　The　comparison　with　MCS　shows　that　this　strategy　has　high　accuracy　and　efficiency.　Probabilistic　structural　response　of　composite　beams　with　different　correlation　length,　coefficient　of　variation　of　interface　stiffness　and　interface　stiffness　of　were　analysis.　The　results　show　that　the　correlation　length　has　little　effect　on　the　probability　distribution　of　the　response　compared　with　the　interlayer　stiffness　and　its　coefficient　of　variation　(COV).　The　COV　of　interlayer　stiffness　has　little　influence　on　the　mean　value　of　the　response　but　has　great　influence　on　the　dispersion　of　the　result.　The　greater　the　COV　is,　the　greater　the　dispersion　of　the　result　is.　The　interlayer　stiffness　has　a　great　influence　on　the　statistical　characteristics　of　the　response,　and　within　a　certain　range,　the　greater　the　interlayer　stiffness,　the　smaller　the　deformation　of　the　composite　beam　and　the　smaller　the　probability　range.