After many years of studying of nanofluids development, their thermophysical properties are not yet known precisely and the judgment of their true potential is difficult. This fact was demonstrated by studying numerically forced convection heat transfer at high Re of a nanofluid consisting of water and Al 2 O 3 in horizontal tubes. Five different models from the literature are used to express the thermophysical properties in terms of particle loading and they led to different results in heat transfer enhancement. In particular, the heat transfer coefficient of water-based Al 2 O 3 nanofluids is increased by 2.33-26.45 % under fixed high Reynolds number compared with that of pure water. As a conclusion, the effect of uncertainties in adopting nanofluids properties was observed.