The angiotensin-converting enzyme 2 (ACE2) and main protease (MPro), are the putative drug candidates for the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). In this study, we performed 3D-QSAR pharmacophore modeling and screened 1264479 ligands gathered from Pubchem and Zinc databases. Following the calculation of the ADMET properties, molecular docking was carried out. Moreover, de novo ligand design was performed. MD simulation was then applied to survey the behavior of the ligand-protein complexes. Furthermore, MMPBSA was utilized to re-estimate the binding affinities. Then, a free energy landscape was used to find the most stable conformation of the complexes. Finally, the hybrid QM-MM method was carried out for the precise calculation of the energies. The Hypo1 pharmacophore model was selected as the best model. Our docking results indicate that the compounds ZINC12562757 and 112260215 were the best potential inhibitors of the ACE2 and MPro, respectively. Furthermore, the Evo_1 compound enjoys the highest docking energy among the designed de novo ligands. Results of RMSD, RMSF, H-bond and DSSP analyses have demonstrated that the lead compounds preserve the stability of the complexes’ conformation during the MD simulation. MMPBSA data confirmed the molecular docking results. The results of QM-MM showed that Evo_1 has a stronger potential for specific inhibition of MPro, as compared to the 112260215 compound. In conclusion, our results showed that the de novo designed Evo_1 compound has the potential to be used as a drug for the treatment of COVID-19; however, further in vitro and in vivo validations are required.