The development of drug-resistant viruses limits the therapeutic success of anti-HIV therapies. Some of these genetic HIV-variants display complex mutational patterns in their pol gene that codes for protease and reverse transcriptase, the most investigated molecular targets for antiretroviral therapy. In this paper, we present a computational structure-based approach to predict the resistance of a HIV-1 protease strain to amprenavir by calculating the interaction energy of the drug with HIV-1 protease. By considering the interaction energy per residue, we can identify what residue mutations contribute to drug-resistance. This approach is presented here as a structure-based tool for the prediction of resistance of HIV-1 protease toward amprenavir, with a view to use the drug-protein interaction-energy pattern in a lead-optimization procedure for the discovery of new anti-HIV drugs.