This article presents results of an ongoing investigation into the modeling of thermal conductivity for Alumina-Water nanofluids. In spite of having the promise of being an improved heat transfer medium, fundamental understanding and modeling of important thermo-physical properties of nanofluids (such as thermal conductivity) have remained a difficult task due to the possible influence of several particle and base-fluid properties on the behavior of nanofluids. The existing theories to explain the phenomenon of thermal conductivity augmentation have provided different and sometimes contrasting mechanisms. In this study, seven existing theoretical models for thermal conductivity of nanofluids have been evaluated for their accuracy by comparing the predicted versus experimental data for a wide range of test conditions. The existing models were found to provide inaccuracies (over/underpredictions) in the range of 3 to 58%. A new model has been developed using dimensionless analysis, which includes Prandtl number and a new dimensionless number that is a ratio of Reynolds number to the square root of Brinkman number for particle and fluids. The new model has been found to generally predict the thermal conductivity ratio (nanofluids to base fluids) within 5% accuracy range.