In this work, we characterize the interaction of angiotensin-I-converting enzyme (ACE) inhibitors with their receptor derived from the Binding Database by combining ligand-based and structurebased methods. The ligand-based quantitative structure-activity relationship (QSAR) model by support vector machine (SVM) achieves an overall accuracy of 88.74%, Matthews correlation coefficient of 0.678, and area under the receiver operating characteristic curve of 0.914 with leaveone-out (LOO) cross-validation on 444 training samples. The predictive ability of the model obtained is further verified by predictions on two test sets including 110 and 114 compounds. We show that the SVM-based model, with 2D and 3D QSAR advantages, is simple, accurate, and robust and can be used to predict and identify new ACE inhibitors. The four descriptors, namely the capacity factor, volume, standard deviation, and hydrophilic-lipophilic features, in the QSAR model can well represent the SAR of these inhibitors. In parallel, the structure-based molecular docking studies reveal that hydrogen bond is an important force for the binding affinities of the ACE inhibitors with the receptor. This work is useful in understanding the interaction mechanisms of ACE inhibitors with their receptor, as well as designing of new ACE inhibitors.