Drug repurposing is an active area of research and effort due to decreasing the cost and time of drug development. Most of those efforts are primarily concerned with the prediction of drug-target interactions. Many evaluation models, from matrix factorization to more cutting-edge deep neural networks, have come to the scene to predict drug-target interactions. Most of the available information on drugs and targets is gathered and used as the features to feed the prediction models. Some predictive models are devoted to the prediction's quality, and others are devoted to the efficiency of the predictive models, e.g., embedding generation. In this work, we propose two predictive models of drug-target. To do this, we use the relations of drugs and targets and propose a method of similarity computations. Using these similarities, we generate an accumulative feature representation of these two objects. We propose two inductive, deep network models of IEDTI and DEDTI for drug-target interaction prediction. The former uses triplet and maps the input feature vectors into meaningful embedding vectors. Then, it applies a deep predictive model to each drug-target pair to evaluate their interaction. The DEDTI directly uses the feature vectors of drugs and targets and applies a predictive model to each pair to predict their interactions. The results show that both models outperform the state-of-the-art models.