The heat transfer and flow of nanofluids in microchannel heat sink is usually numerically investigated by the single-phase model. The thermal conductivity and viscosity of nanofluids are calculated in the whole computational region because the single-phase model, assumes the nanofluids is static and nanoparticles concentration is uniform. However for the actual forced flow process, the concentration distribution of the nanofluids is not uniform. So the thermal properties of the nanofluids vary within different computational regions. With the increase of the velocity of the nanofluids, the deviation between the prediction of single-phase model and experimental results increases. In this paper, a new simulation method is proposed to consider the inhomogeneous and dynamic thermophysical properties of the nanofluids by obtaining the nonuniformly distributed nanoparticle concentration in each numerical cell. At the same time, it takes into account the following factors: the interactions between the nanoparticles, the solid wall and the base fluid, the axial conduction, the slip boundary conditions and fluid-solid coupling. The results show that the proposed numerical simulation method further improve the simulation accuracy for the forced flow and heat transfer process of nanofluids in the microchannel in a wider range. The effects of the nanoparticles concentration, Reynolds number and axial heat transfer on the thermophysical properties of the nanofluilds, the flow and heat transfer characteristics of a rectangular microchannels with hydraulic dimensions of 341μm were analyzed using the proposed simulation model. The main results are as follows: the nanoparticles are enriched at the near wall for the fully developed region; the nonuniformity of the nanofluilds thermal conductivity coefficient distribution at the channel cross section increases with the increase of the mass flow rate and the inlet centration; under the same Reynolds number, the higher the nanofluids concentration the better the heat transfer characteristics, but there is no linear relationship between the enhancement amplitude and concentration; the heat transfer coefficient of nanofluid with 1% and 2% alumina nanoparticles concentration is about 5.86% and 8.49% higher than that of deionized water, respectively; the axial conduction effect cannot be ignored for the microchannel, which significantly affects the inlet and outlet region of the microchannel especially when the Reynolds number is low. The nanofluilds temperature in the microchannel increases with the curve and does not match the conventional linear growth hypothesis.