Nowadays every manufacturing and industrial industry has to focus on the manufacturing of quality products. Manufacturing of these kinds product with higher accuracy, better surface finish, lower maintenance and lower process planning and manufacturing cost are very important factor that can achieved by using non-conventional optimization techniques instead of conventional techniques. Many of non-conventional optimization techniques like Fuzzy Logic approach based technique, Genetic algorithms, Artificial Neural Network, Particle Swarm optimization, Ant colony optimization, Scatter search technique and simulated Annealing etc. are used to optimization of surface roughness. Milling is a machining operation in which workpiece is fed below the cylindrical rotating multi point cutting tool, multi point cutting tool having multiple cutting edges. On the basis of literature review, many machining parameters such as cutting speed, feed rate, depth of cut, cutting fluid pressure etc. and performance parameters as surface roughness, material removal rate, tool wear ratio, tool vibration etc., were observed for CNC milling operation. The correct selection of machining parameters is very important factor to achieve best performance measure. In this research work, spindle speed (SS), feed rate (FR) and depth of cut (DOC) are selected as machining parameters while surface roughness is considered as performance parameters to perform end milling operation on the workpiece materials of 6101 Aluminum alloy, Copper of electrolytic grade and Mild Steel 2062 by using High Speed Steel (HSS) end mill cutter of 12 mm diameter. Minimum experiment trials are designed by Taguchi based L9 (3^3) orthogonal array with the help of Minitab 17.0 software and a fuzzy logic approach based model is taken as to predict the value of surface roughness of a machined surface in 6101 aluminum alloy, Copper of Electrolytic grade and Mild Steel 2062 milling operation using HSS end mill cutter of 12 mmdiameter. Three membership functions are allocated to be connected with each input of the model. The predicted results achieved via fuzzy logic model are compared to the experimental result. The result demonstrated settlement between the fuzzy model and experimental results with the 95.618% model accuracy for 6101 aluminum alloy material, 83.849% for copper (Electrolytic grade) and 98.334% Mild Steel 2062.