According to the acoustic emission signals obtained from the simulation of the holding-load and load servo cylinder of the rolling automated gauge control (AGC), this paper presents a hits-density based visualization model for early-stage vibration state identification of the AGC hydraulic cylinder. First, the normal loading and early-stage vibration are applied to a 660mm rolling mill AGC cylinder to acquire the acoustic emission bursts. Then, particular signals are extracted according to their cumulative statistical distribution analysis to describe the changes of the servo cylinder under load. Finally, the acoustic emission hits density images of the ACG servo cylinder under normal loading and with early-stage vibration are drawn after special organization of the selected features and some logical calculation with limits in range. The hits density images suggest promising performance of such images for the early vibration identification of AGC servo cylinders.