False (artificial) diffusion provides an erroneous estimation of molecular diffusion during the simulation of liquid micromixing. The present review introduces discretization methods, numerical grid types, and numerical errors to address the effect of false diffusion on the prediction of mixing efficiency of microfluidic devices. False diffusivity is characterized by the grid Peclet number, grid type, and discretization scheme. This review demonstrates that most investigators have selected a grid resolution just for the grid independence test. It is revealed that the convergence criterion should not be quantitative values of mixing efficiency even when high-order schemes are employed to discretize the computational grid. Based on the previous publications in this field, a straightforward procedure is recommended to manage false diffusion in the numerical simulation of micromixers.