Purpose: Subtle myocardial deformations such as early systolic lengthening (ESL) and post-systolic shortening (PSS) seem to be useful for detecting myocardial ischemia by 2-dimensional (2D) speckle tracking echocardiography. However, it is still unclear whether the same holds true for the 3-dimensional (3D) method. We therefore evaluated the accuracy of 2D and 3D speckle tracking echocardiography for the diagnosis of myocardial ischemia using an experimental model. Subjects and Methods: In 20 open-chest dogs, 2D or 3D echocardiographic images were acquired before flow reduction of the left circumflex artery, during 20-40% of flow reduction (stenosis 1) and 40-60% of flow reduction (stenosis 2). Peak systolic strain, early systolic lengthening, and post-systolic shortening were analyzed in the ischemic region, and the area under the curve (AUC) of the receiver operating characteristic curve for diagnosing each stenosis was calculated. Results: With the 2D method, the AUC of all parameters was significantly larger than the line of no information (AUC=0.5) in the case of stenosis 2; however, only the AUC of PSS was significant in the case of stenosis 1 (AUC: peak systolic strain=0.70, ESL=0.72, PSS=0.87). With the 3D method, the AUC of all parameters tended to be smaller than that yielded by the 2D method. Neither ESL nor PSS was significant in the case of stenosis 1 (AUC: peak systolic strain=0.64, ESL=0.51, PSS=0.61). Conclusion: 2D speckle tracking echocardiography provided better diagnostic accuracy for detecting myocardial ischemia than the 3D method. The assessment of PSS was superior to that of peak systolic strain with the 2D method but not with the 3D method.