To　overcome　the　difficulties　of　human　reliability　analysis　in　risk　assessment　for　shield　tunneling,　this　paper　proposes　a　weighted　fuzzy　cognitive　reliability　and　error　analysis　method(CREAM)model.　First,　the　reasonable　weight　of　common　performance　conditions(CPCs)is　obtained　by　multiple　correlation　analysis　and　evidence　theory.　Second,　the　specification　evaluation　rules　for　CPCs　are　established　based　on　the　current　standard,　and　the　membership　function　for　each　level　of　CPCs　is　obtained　by　statistical　method.　On　this　basis,　a　continuum　technique　for　the　discrete　basic　diagram　for　control　mode　is　proposed,　and　an　estimate　model,　which　can　be　used　to　determine　the　human　error　probability(HEP)in　the　construction　of　shield-driven　tunnel,　is　established.　Then,　the　rationality　and　reliability　of　the　model　are　verified　by　four　axioms.　Results　show　that　the　method　of　introducing　the　weight　into　the　calculation　of　the　expected　CPC　effect　and　the　HEP　reasoning　in　the　weighted　fuzzy　CREAM　model　can　better　reflect　the　effect　of　CPCs　on　human　performance.　This　model　can　also　provide　a　reference　for　quantitative　HRA　in　shield　tunneling　of　subways.　©　2019,　Editorial　Board　of　Journal　of　Tianjin　University(Science　and　Technology).　All　right　reserved.