Prediction of drug metabolism is an important step in the early lead optimization phase and preclinical studies. Its main purpose is to avoid late stage (Phase I–III) or post‐marketing withdrawals which usually results in a significant financial loss to a pharmaceutical company. Computational methods for identification of soft‐spots in the molecules (site of metabolism; SOM) and CYP450 isoforms responsible for drug metabolism have over the past ten years proved an indispensable tool towards achieving this goal. These methods can be broadly classified into i) ligand based, ii) structure based and iii) hybrid methods. In this review we discuss some of the most successful methods of SOM and isoform specificity prediction, their application domain, advantages, and limitations along with recent developments in the last 5 years in this area. Recent applications of molecular dynamics (MD), quantum mechanics (QM) and quantum mechanics/molecular mechanics (QM/MM) methodology for regioselective and isoform specific drug metabolism predictions are also discussed. Finally, a summary of these methods along with expected developments, future challenges and opportunities are discussed.