The evaluation of multi-scale surface roughness parameters (SRPs) is important to solve many engineering problems (e.g. contact stress, sealing, friction) and is closely related to further fundamental problems (e.g. microbial contamination). Traditionally, surface roughness has been used as a standard for indicating process performance, such as tool wear, tool vibration etc. This paper also aims to find appropriate surface roughness parameters (SRPs) that can be used as process monitoring indices. Grade 304 stainless steel surfaces, generated by extrusion and grinding processes, were used in this study. The evaluation of different SRPs and their topography properties (such as fractal dimension) is discussed for extruded and ground surfaces. One problem with existing surface metrology is the availability of a multitude of disconnected roughness parameters. A statistical approach is presented in this paper that allows the most appropriate roughness parameters to monitor whether the intended surface quality converges to be found.