In this paper, a detailed novel approach is presented to adequately predict the root mean square surface roughness parameter (Rq) in electro-discharge machining (EDM) process by taking into account the effects of consecutive discharges and plasma flushing efficiency (PFE). The procedure first starts simulating crater formation due to a single spark based on more realistic assumptions using finite element method (FEM). The Abaqus FEM software has been employed to estimate the transient temperature distribution, crater cavity dimensions and volume to obtain the removal rate per discharge pulse on AISI 304 stainless steel as the work piece. To evaluate the amounts of (PFEs), the theoretically ideal material removal rates (MRRs) were compared with a set of experimental MRRs under the same low discharge energy zone (less than 100mJ) typically used for finishing purposes. Finally, an iterative-based statistical approach derived from the concept of normal distribution of surface heights along with the impacts of each crater cavity formed due to repetitive discharges on the surface profile have properly been formulated to simulate the final work piece surface roughness achieved under the erosive effects of multiple discharges. The simulated values of Rq have also been validated experimentally. Reasonable agreement has been attained which confirms the robustness and correctness of proposed methodology.