Fuzzy mathematics is used to produce computer models directly from engineering data acquired from mineral processing systems. A fuzzy model is developed that uses linguistic terms to describe the relationships between the variables in select mineral processing systems. Such fuzzy linguistic models rely exlusively on data collected from the systems being modelled; they do not presuppose an understanding of the physics of the system. The lack of requirement for fully understanding the physics of the system being modelled is important in the mineral processing industry where these relationships are complex and difficult to acquire. The linguistic computer models that are developed are tuned with a genetic algorithm which is employed to simultaneously select linguistic rules and tune fuzzy membership functions. The result is an extremely flexible approach to modelling that is applicable in a variety of mineral processing systems. The effectiveness of this approach to meddling is demonstrated using data from two pieces of mineral processing equipment: (I) a grinding circuit and (2) a hydrocyclone.