In this study, classical and vortex tube cooling methods are compared in the pocket machining of AA5083-H36 alloy with uncoated cemented carbide cutting tool. The effects of cutting speed, feed rate, axial/radial depth of cut and nose radius and their two-way interactions on the surface roughness, and the optimization of surface roughness are investigated via Taguchi method. The experiments conducted based on Taguchi's L16 orthogonal array (OA) are assessed with analysis of variance (ANOVA) and signal to noise (S/N) ratio. As a result, in both cooling methods, it is obtained that roughness correlates negatively with cutting speed and radial depth of cut and positively with feed rate and axial depth of cut. While in the cooling with vortex tube, lower average R (a) values are observed in the experiments with the nose radius of 0.8 mm, in the classical cooling almost no change is obtained. Lastly, optimum roughnesses for the classical and vortex tube cooling are obtained as 0.164 and 0.188 mu m, respectively.