Methods such as fuzzy logic and artificial neural networks have been used frequently recently in modeling. In this study, mathematical models for estimating surface roughness were created in surface milling processes using artificial intelligence techniques. Within the scope of the study, 1.2738 mold steel was used as workpiece material. The experiments were carried out under dry cutting conditions and using a minimum quantity lubrication technique. As cutting parameters, the cutting speeds and the feed rates are selected. In experiments; 80 mm/min, 130 mm/min, 180 mm/min values, 3 different cutting speeds and 0.5 mm/teeth, 0.8 mm/teeth and 1mm/teeth 3 different feed rates were used. A total of 27 experiments have been conducted and the results of the experiments were modelled using the MATLAB program and the effect of the minimum quantity lubrication (MMY) technique on surface roughness using an adaptive network-based fuzzy logic inference system (ANFIS) approach. At the same time, the mathematical model was created by performing regression analysis in MINITAB. The results obtained by ANFIS and regression analysis were compared. As a result, the ANFIS model provided 100% accuracy, while the regression model achieved 71% accuracy.