Most commercial servo systems utilize cascaded proportional–integral–derivative controller due to its simplicity and robustness. Under this control structure, motion performances such as single-axial tracking and multi-axial contour following are completely determined by the controller parameters. In real applications, current loop controller parameters usually remain unchanged while velocity loop and position loop ones should be tuned as the mechanisms including inertia and resonant frequency vary. This article focuses on such cascaded proportional–integral–derivative controller parameters tuning for contour following performance improvement. It first describes the servo system dynamics under the cascaded proportional–integral–derivative control structure and identifies the model parameters via relay feedback technology. Then, the effects of the closed-loop dynamics on contour following are analyzed to gain that the matched axial dynamics is preferable. In order to match the axial dynamics, the controller parameters are tuned loop by loop through setting all the axes to have the same bandwidths. Based on the identified results, feedforward controller parameters including nonlinear coulomb friction compensation are also tuned to further improve the contour following performance. At last, the proposed tuning method is verified through the ellipse and diamond following experiments carried out on an X-Y motion stage. The results show that it can match the axial dynamics effectively and thus reduce the contour error greatly.