Computational approaches are becoming more popular in pharmaceutical development due to their hurdles, particularly in drug-target interaction prediction. Beyond binary classification, evaluating the strength of these interactions, known as Drug-Target Binding Affinities, provides more information. This article examines current computational approaches to forecast the binding affinities between drugs and targets, with a focus on artificial intelligence, machine learning, and deep learning. A thorough comparative analysis, based on a systematic literature review, investigates their differences, strengths, and limits. The performance of different computational approaches for predicting the affinity for drug-target binding implies their performance, applicability, and shifting patterns. These themselves form part of the unending endeavor toward the enhancement of effectiveness and success within pharmaceutical research. Results from this study will thus offer guidelines on how various computational tools may be selected and used in the course of drug discovery.