Steel plated rolling is an important part of steel production, which is composed of multiple processes. Effective matching, tracking and analysis of the plate thickness data is of great significance to the production efficiency and quality detection of steel plate. Aiming at the problem that the rolling speed and sampling frequency of each rolling process are dynamically changed, an improved Dynamic Time Warping (DTW) algorithm with adaptive scaling is proposed in this paper. This method calculates the relative offset and its direction of two non-equal length sequences by using the warping distance matrix, and then adaptive scaling is performed on the heads and ends of the sequences to find the best matching position of the query sequence. In addition, as the length of the steel plate data is too large, the method segments the long main sequence to perform adaptive scaling efficiently, which significantly reduces the time consumption. The simulation experiments on real industrial data showed that the proposed method could find the reasonable matching positions and effectively decrease the computing time complexity, which laid the foundation for the tracking and analysis work of the rolling process.