As a significant similarity measure, dynamic time warping has critical effect on the performance of time series data mining. The study of dynamic time warping has significant theoretical and practical importance. Firstly, the basic step of the dynamic time warping algorithm is introduced, and the advantages and the shortages are analyzed. Then, dynamic time warping at the aspects of the related improved research of measure efficiency, promotion ways of measure quality and its application in various industries is reviewed. Furthermore, relative some research hotspots are summarized, and several directions for the future research is given. The review and analysis for the related methods of dynamic time warping can provide necessary document literature and theory basis for time series data mining including similarity, clustering and classification.