With the increasing miniaturization and precision requirements for products, the Micro Electro Discharge Machining (hereinafter Micro-EDM) has been world-widely concerned for advantages as high efficiency and good processing quality. Assessment parameters and methods of machined surface topography are commonly adapted in the Micro-EDM surface roughness assessment that blocks the further development of the Micro-EDM technology. This research improved the traditional Gauss filtering technique by Micro-EDM technology and morphological characteristics, meanwhile extracted the 2D and 3D surface roughness ion and carries out simulation experiments of the algorithm. Simulation results show the improved Gauss filtering algorithm is adapted to the Micro-EDM surface roughness extraction for all different sample data.