Stochastic　dynamic　analysis　of　offshore　structures　considering　both　fluid-structure　interaction　and　system　uncertainties　is　a　challenging　research　task.　This　paper　proposes　a　novel　method　to　evaluate　the　statistical　characteristics　of　offshore　structural　responses　with　consideration　of　uncertainties　in　the　fluid　and　structure.　The　fluid　behavior　is　simulated　by　a　finite　number　of　particles　with　particle　finite　element　method　(PFEM),　and　the　dynamic　behavior　of　the　offshore　structure　is　modelled　with　finite　element　method　(FEM).　A　PFEM-FEM　scheme　is　used　to　model　the　fluid-structure　interaction.　To　evaluate　the　statistical　characteristics,　spectral　representations　method　is　used　for　uncertainty　propagation.　The　output　of　fluid　particles　position/pressure,　structural　vibration,　etc.　are　represented　by　using　polynomial　chaos　(PC)　expansion,　and　their　coefficients　are　obtained　from　the　least　squares　method.　Statistical　characteristics　of　the　responses,　such　as　mean　value　and　variance,　can　be　evaluated　with　the　obtained　PC　coefficients.　Three　numerical　examples　are　studied　in　this　paper.　The　first　example　is　a　simple　structural　model,　which　is　used　to　demonstrate　the　convergence　and　accuracy　of　the　uncertainty　analysis　method.　The　second　example　is　a　benchmark　dam　break　problem.　Statistical　characteristics　of　the　fluid　particles　position　due　to　uncertainties　in　the　mass　density　are　evaluated.　Numerical　results　are　verified　with　experimental　data　and　observations.　In　the　third　example,　PFEM-FEM　scheme　is　used　to　conduct　fluid-　structure　interaction　analysis.　The　marine　riser　structure　is　modelled　with　beam　elements.　In　the　fluid　domain,　PFEM　is　used.　Uncertainties　in　both　the　fluid　domain　and　structural　domain　are　considered.　Results　demonstrate　that　the　proposed　approach　can　be　used　to　evaluate　the　statistical　characteristics　of　responses　in　the　fluid-structure　interaction　analysis　accurately　and　efficiently.