The smoking process is a batch one with long duration processing (some hours). Deposition of smoke particles can be enhanced by electrostatic precipitation. However, a better knowledge of the fundamental physics is needed to optimize the process. We carried out experimental and numerical studies to obtain essential information on the smoke flow. First, experiments physically characterized the wood smoke. These measures lead us to study the turbulent two-phase flow using an algebraic slip model. The main coupling with electrical forces is taken into account through the slip velocity. Various assumptions were tested concerning the coupled phenomena in order to build an algorithm that is able to simulate these phenomena. Numerical resolutions are compared with data from the literature on a simple electrostatic precipitator (three wires between two plates). Simultaneously, experiments were performed on a simplified scaled-down model. of the smoking process (wires and one plate). The results of particle image velocimetry show clearly that the electrostatic precipitation phenomenon is effective above a threshold voltage and is significantly influenced by geometry of electrodes. Experiments and numerical resolutions underwent a validation phase. A small-size industrial demonstrator, of which the promising results will not be discussed in this paper, was conceived by exploiting the presented works.