Abstract: Time series classification is an important topic in data mining. Most existing time series classification methods do not consider the influence of the shape of the time series on the classification results. The paper proposes a fuzzy classification method for time series based on k-shape. The method utilizes the k-shape clustering algorithm to cluster each category of the time series training datasets and obtains the cluster centers group of each class. After utilizing the cluster center group of each class as the initial clustering center of the fuzzy classification, class labels of the test datasets are determined according to the principle of maximum membership degree. Experimental results on 30 time series public datasets show that the proposed method has better classification performance than the traditional methods, including support vector machine (SVM), Bayes, ensemble algorithm of interval weightsc (EAIW), and trend information based on longest common subsequence (TLCS), with more excellent usability for time series with distortion and displacement characteristics.