The need to find new antibiotics is expanding as a result of the quick rise in bacteria that are resistant to medicines. Discovering drug-protein interactions could be an essential first step in the process of developing drugs since it will substantially reduce the scope of the look for possible solutions. Since in vitro assays are extremely time-consuming and pricey. We developed a machine learning method that can predict medications for the target in order to overcome this difficulty. We used the Padel script to do predictions on several chemical libraries, acquire drug physical and chemical properties, and obtain features extracted. establishing which model is best for predicting drug-target interactions is performed by analyzing the Random Forest technique with the Naive Bayes method, K-Nearest Neighbor, and other choices. This study reduces the failure rates and costs incurred when creating new pharmaceuticals while demonstrating the value of adopting machine learning approaches in drug discovery.