Data　from　full-scale　building　tests　conducted　in　Texas　Tech　University　are　used　to　quantify　the　uncertainties　associated　with　the　use　of　computational　fluid　dynamics　to　obtain　wind　load　predictions　for　full-scale　structures.　It　is　demonstrated　that　the　effects　of　computational　domain　is　relatively　small,　and　the　effects　of　grid　arrangement,　turbulence　model　and　inlet　boundary　condition　are　relatively　obvious.　Finally,　guidelines　are　suggested　for　reasonable　design　and　suitable　choice　of　turbulence　model.