Aim/Background: The anomalous production of dihydrotestosterone (DHT) observed in benign prostatic hyperplasia and prostate cancer particularly in tissues of the prostate gland. The primary male sex hormone, testosterone (T) is improved to a metabolite in cells by 5a-reductase (5aR) type II enzyme through NADPH. A metabolite is DHT which is more potent than T. Drug therapy is the best alternative for the treatment of benign prostatic hyperplasia (BPH). Drug therapy act by dropping the DHT formation through inhibiting the 5aR enzyme. Thus, this research work was undertaken to design a novel 5aR enzyme inhibitor by pharmacophore and Atom-based 3D QSAR technique. Materials and Methods: A dataset of twenty-nine ligands available with IC 50 chosen from the literature. Schrodinger molecular modelling software is having, phase 3.0 module implicated for generation of pharmacophore models. Results: Pharmacophore hypothesis with five features having two H-bond acceptors and three hydrophobic group was developed, i.e., AAHHH.715. This Pharmacophore hypothesis regarded as the finest one. This hypothesis resulted into statistically significant three-dimensional QSAR model. The statistical parameters were found to be 0.9804 as r2 value and 0.8321as q2 value. Out of 29 ligands, 23 ligands assigned as training set and 6 ligands as a test set. The squared correlation coefficient between training and test sets based on actual and predicted values were observed to be 0.96 and 0.87 respectively. Conclusion: The 5aR enzyme inhibitors predicting requirements can be done by this built model.