Environment　protection　is　important　for　the　survival　of　residents,　and　the　government　must　improve　its　governance　model　on　environmental　cost　prediction　methods　to　address　the　increasing　level　of　environmental　pollution.　Therefore,　a　science-based　investment　scheme　is　of　great　significance.　To　improve　the　accuracy　and　effectiveness　of　environmental　governance　cost　prediction　method,　it　is　important　to　consider　the　completeness　of　the　indicators　and　their　degree　of　contributions-both　of　which　need　to　be　studied　further.　Considering　the　influence　of　a　decision-maker＇s　subjectivity　on　an　investment　scheme,　this　paper　proposes　a　prediction　method　accommodating　the　risk　preferences　of　different　decision-makers.　The　proposed　method　is　based　on　the　synthesis　of　evidential　reasoning　approach.　An　objective　empowerment　is　carried　out　according　to　the　standard　deviation　method　of　correlation　coefficient　to　highlight　the　importance　degree　of　different　indicators.　At　the　same　time,　to　improve　the　practical　usage　of　the　synthetic　cost　prediction　method,　the　future　cost　is　predicted　by　combining　the　genetic　programming　models　under　different　risk　coefficients,　namely,　the　risk　preference,　the　risk　neutrality,　and　the　risk　aversion.　Finally,　a　case　study　involving　environmental　governance　cost　prediction　of　29　provinces　of　China　is　presented.　A　comparison　of　the　cost　predictions　and　the　actual　value　of　different　risk　coefficients　for　the　different　methods　are　given　to　evaluate　the　effectiveness　of　the　proposed　method.　©　2018　Elsevier　Ltd