Abstract: Modeling and diagnosing using MPCA/MPLS will lose efficiency because of different time length of data trajectories in batch process fault diagnosis. The Dynamic Time Warping (DTW) is used to synchronize the time length of data trajectories observed.Both improved DTW, recursive and coarse-grid points, are proposed. The simulations show that the two methods can not only synchronize data trajectories successfully, but also be applied to on-line fault diagnosis.