High speed machining is known as an advanced machining process increasingly used for modern materials such as nickel or cobalt based superalloys. Besides increasing productivity and accuracy, high cutting speed will also improve surface quality. This paper outlines modeling methodology applied to optimize cutting parameters during high speed milling in order to obtain a very good surface finish. The parameters taken into account were cutting speed, cutting feed and axial depth of the cut. These parameters are easy to controll and improvements can be made without costs, simply by setting them at optimal values. For economic reasons, in order to solve the problem with minimum number of trials, a centered composite experiment design was used. Experiment results of surface roughness measurements after high speed end milling of cobalt based superalloy FSX 414 with Ti coated carbide tools are presented. A predictive model was obtained after analyzing with ANOVA the influence of cutting parameters on surface roughness. The predictive model-well correlated with experimental data-shows that surface roughness does not depend on axial depth and is mostly influenced by feed rate. Cutting speed must be increased and cutting feed decreased to obtain a lower surface roughness Ra.