Initial screened com-pounds were passed through several criteria including the range ofactivity of training set, fit to model with omitted feature set to ‘0’and comparison of dock score and binding energy with most activecompound of the data set to reach the potential compounds. Amongthe more than twenty five thousand initial hit compounds, sevencompounds finally satisfied all these criteria and were used to fur-ther evaluate the binding interactions with critical amino residuesat the active site of HIV protease. This virtual screening usingthe Hypo 1 obtained one potential molecule, NCS70804 from NCIdatabase which is already confirmed as active in anti-HIV screening.A number of hydrogen bonds and bump interactions were observedbetween potential HIV-protease compounds and catalytic residuessuch as Leu24, Asp25, Thr26, Gly27, Asp29, Ile50 and Ile84. There-fore this indirect receptor-independent pharmacophore model cancalculate accurately the position and orientation of potential ligandin a binding site and the selected model has enough potentialto retrieve the active molecule from database. The potential HIV-protease molecules will be subjected to experimental validation toobtain further confirmation.