This　paper　presents　an　investigation　on　prediction　accuracy　of　several　two-equation　turbulence　models　on　gas　turbine　stator　and　rotor　blade　tip　and　casing　heat　transfer.　The　prediction　ability　of　SST　γ-θ　model　on　heat　transfer　at　varying　turbulence　intensities,　turbulence　integral　length　scales　and　Reynolds　numbers　are　also　studied.　The　results　indicate　that　the　SST　γ-θ　model　has　significantly　improved　the　prediction　accuracy　on　stationary　cascade　surface　heat　transfer.　The　numerical　results　of　SST　γ-θ　model　show　well　agreements　with　the　experimental　data　at　a　large　range　of　inow　turbulence　intensities　and　exit　Reynolds　numbers.　The　predicted　heat　transfer　coEfficients　on　blade　tip　and　casing　by　using　different　turbulence　models　present　similar　distribution　trends　with　the　experimental　data,　however,　the　results　of　SST　γ-θ　model　show　best　agreements　on　blade　tip　and　the　most　regions　of　casing　surface.　©　2015,　Science　Press.　All　right　reserved.