Human glutaminyl cyclase (hQC) appeared as a promising new target with its inhibitors attracted much attention for the treatment of Alzheimer's disease (AD) in recent years. But so far, only a few compounds have been reported as hQC inhibitors. To find novel and potent hQC inhibitors, a high-specificity ZBG (zinc-binding groups)-based pharmacophore model comprising customized ZBG feature was first generated using HipHop algorithm in Discovery Studio software for screening out hQC inhibitors from the SPECS database. After purification by docking studies and drug-like ADMET properties filters, four potential hit compounds were retrieved. Subsequently, these hit compounds were subjected to 30-ns molecular dynamic (MD) simulations to explore their binding modes at the active side of hQC. MD simulations demonstrated that these hit compounds formed a chelating interaction with the zinc ion, which was consistent with the finding that the electrostatic interaction was the major driving force for binding to hQC confirmed with MMPBSA energy decomposition. Higher binding affinities of these compounds were also verified by the binding free energy calculations comparing with the references. Thus, these identified compounds might be potential hQC candidates and could be used for further investigation.