<p>Monitoring part quality is essential in stamping for making timely production decisions to avoid defects. In this context, a process energy map in stamping, which indicates the accumulation of the force that deforms the blank over stamping depths, is proposed to visualize and monitor part thickness variation by color and its intensity. The stamping depth and process energy are set as the horizontal and vertical coordinates of the map, respectively. A construction strategy using stacking machine learning (SML) for a thickness variation prediction model is developed to generate a large amount of accurate thickness variation data to fill in the map. Correlation analysis is included in the strategy to select the machine learning models that are integrated into the prediction model. Then, the map is colored according to the data value, followed by the quality zone division in the map based on the threshold curve of thickness variation. A downscaling part of a car door was formed to validate the effectiveness of the proposed map. The mean absolute percentage error and mean square error of monitored thickness variation at different stamping depths were within 5.03% and 0.94%², respectively, showing the high accuracy and stability of the built map. The results of applying the map in stamping showed a 14.56% reduction in the maximum thinning ratio of the part, which effectively prevents cracking. The proposed process energy map assisted in accurate part quality monitoring and quality improvement in sheet metal stamping.</p>