The Soret-effect or thermo-diffusion in liquids has been a subject of long-standing consideration for research. Numerical modelling of thermophoresis in water-alumina nanofluid subjected to sub-cooled pool boiling conditions is attempted in this paper. Pool boiling experiments are conducted using water-alumina nanofluid for different heater input conditions (50 °C–90 °C). Computational fluid dynamics (CFD) code ANSYS Fluent is used further to analyse thermophoresis in water-alumina nanofluid. 2-D steady pressure-based implicit solver with Gauss-Seidel iterative procedure is used to discretize and numerically solve the mathematical model. Pressure-correction in the numerical procedure is done using SIMPLE (Semi-implicit Method for Pressure Linked Equation) technique. Flow boiling models for micro-sized particles are revised to outfit pool boiling conditions of nanofluid. Results of temperature from the numerical model were found to have a worthy-match with measured temperature values of the experiments. Uncertainty analysis is carried out for the random variation of temperature measurements in experiments. Thermophoretic parameters, Peclet numbers (rotational/translational) and time-scale equations for diffusion are used to analyse minute particle motion due to advection/diffusion nano-scale heat transfer. Results of thermophoretic velocity reveal a diminishing effect of particle motion near heater surface with amplified heater inputs. Results of Peclet numbers (translational/rotational) reveal the significant role of very small nano particle motion on heat diffusion in nanofluids. Combined results of Peclet number and time-scale equations reveal the strengths of advection and diffusion mechanisms in the nano-scale heat transfer process. © 2018 Elsevier Masson SAS