Drug advancement depends vigorously on the compelling identification of potential interactions among prescriptions and proteins. Finding these relationships is as yet tedious and asset serious, even following quite a while of trial research. Thus, various PC strategies have been created to gauge drug-target correlations for an expansive scope. In this work, we give a profound learning-based way to deal with foresee drug-target interactions dependent just upon protein succession information and drug structures. With exactness paces of up to 92.2% for GPCRs, 90.2% for nuclear receptors, 92.2% for ion channels, and 90.7% for enzymes in our dataset, our discoveries show the viability of our technique. Urgently, on normal benchmark datasets, our model outflanks present status of-the-workmanship computational procedures. Additionally, the findings of our experiments demonstrate the potential of our method to identify tiny yet important characteristics, which makes it a useful tool in the hunt for novel medications.