Abstract: In order to obtain the dynamic cutting force， the experiment of 12Cr18Ni9 revolving parts was carried out on the NC machine tool ETC1625P. The location information of the cutter in the two orthogonal directions of the cutting plane was obtained by the real-time signal acquisition system. Meanwhile， the KISTLER sensor was installed on the tool head of the machine tool to measure the real-time cutting forces in three cutting directions. The experimental data shows that volatility change of the cutting force was produced in the feed direction with the displacement. In order to study the complex nonlinear relationship between the cutting force and cutting parameters， the prediction of dynamic cutting force is realized by time series analysis modeling. The cutting force error is analyzed， and the prediction accuracy obtained meets the requirements.

Key words: machine tool, cutting experiment, dynamic cutting force, time series analysis, predictive modeling