Introduction: Identification of Hypertrophic cardiomyopathy (HCM) morphological subtypes has the same significance as diagnosing HCM due to the close relation of subtypes and genetic mutations. To measure LV wall thickness in echocardiography, i.e. to delineate both endocardiuam and epicardium is not a trivial task. This study aims to automatically classify HCM cases into subtypes via two-dimensional endocardial curvature and dynamic time wrapping (DTW). Methods: Cine MR images were acquired for 19 HCM patients (mean age: 51±13 years, F/M:12/7). All 19 cases were categorized by our senior HCM consultant as sigmoid (n = 6), reverse curvature (n = 8) or neutral (n = 5). Endocardial boundaries were delineated by trained experts on long axis three-chamber images. We chose the polynomial fitting methods to compute endocardial curvature. We used three types of dissimilarity measures: (1) Euclidean distance without alignment, (2) dynamic time wrapping (DTW) distance and (3) soft DTW distance. DTW and its variant soft-DTW were used to quantify the dissimilarity between a pair of endocardial curvature series. The dissimilarity was utlized in classifiers. In our limited data, we used the classic K nearest neighbors method. Results: The average of manual delineation time is half minute. Curvature computation average time is 5 ms on a 2.5 GHz CPU desktop. The average time of DTW for one case is 0.7ms. Soft DTW took about 0.15s for one case. On out-of sample test data, the classifier using Euclidean distance without alignment, DTW and soft-DTW were detected respectively with 52%, 58% and 68% accuracies. Euclidean distance without alignment and DTW have a better in-sample accuracy while soft-DTW has slightly better out-of-sample accuracy. Conclusion: We proposed a method to classify HCM subtypes via softDTW aligning two-dimensional endocardial curvature without delineating the epicardium. The method obtains better classification accuracy than Euclidean distance without alignment in our experiments.