Microemulsion systems may enable innovative and highly efficient synthesis paths for the chemical industry. When expensive catalysts are involved in the process, an efficient separation of products together with an extremely low catalyst leaching are of utmost importance. In order to optimize the separation process and improve the settler design, a thorough understanding of droplet-droplet interactions is required. In this work, computational fluid dynamics (CFD) has been used to study the multiphase flow in a horizontal settler using the Eulerian two-fluid framework. Droplet coalescence, which is the most important physical phenomenon in gravity-driven separation, has been modeled using a spatially inhomogeneous buoyancy-based kernel, and the resulting population balance equation has been solved using the quadrature-based method of moments (QMOM). After calibrating the model parameters using selected experiments, a very good description of the observed separation is obtained, opening the door for process understanding and optimization. (C) 2017 Elsevier Ltd. All rights reserved.