The development of an epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor (TKI) is an ongoing and challenging research field. However, the dynamic motion of the binding site of EGFR has not been accurately depicted, hindering the improvement of EGFR TKI. For this reason, about 33 protein complexes (32 EGFR proteins plus 1 ErbB4 protein) were carefully curated and subsequently studied for dynamic movements of their binding sites by molecular dynamics simulations in this study. The analysis of root mean square deviation (RMSD) revealed that T790M mutation can make an impact on dynamic motion of binding sites; the RMSD value of the EGFR binding site was unrelated to inhibitory activity. The analysis of the radius of gyration (Rg) revealed that T790M can slightly shrink the value of Rg, thereby influencing the shape of the EGFR binding site. More interestingly, the Rg value can exhibit weak correlation with inhibitory activity of most inhibitors. The relationship between Rg and biological activity deserve our serious interest since the best scoring function, Xscore, cannot distinguish highly active EGFR inhibitors. The root mean square fluctuation (RMSF) analysis of key residues derived from binding sites indicated that the most flexible residue was ASP800 with a large RMSF value against the steady residue ALA743 with a small RMSF value, and two other residues (MET793 and LEU844) were supposed to be involved with molecular recognition. In short, the obtained results would be more effective for guiding the development of a novel EGFR kinase inhibitor.