Milling parts for watch, medical, aircraft or molds industries is a compromise between time and precision or surface quality. The latter is very often related to machine vibrations during the process. This paper summarizes results obtained with two approaches aiming at reducing machine vibrations caused by axes accelerations. Both are control model based, taking into account the machine's modal behavior. The first algorithm optimizes the acceleration profile, while the second manages the axes cross-talk vibrations. A high-end 5-axis machine tool was used to mill 8 mm square pockets in brass with a 1 mm diameter tool and using a 0.05 mm depth of cut. Standard CNC parameters as well as vibration reduction options were evaluated. Then the two proposed algorithms were implemented on the same machine-tool using a laboratory Matlab based CNC. This paper describes the significant improvements provided by the algorithms when compared to a high end CNC. In the case of milling reference square pockets, up to 90% vibration amplitude reduction were achieved for a given feed rate, and a 5-fold decrease in the pockets machining time was obtained for the same surface quality.