The family of calcium-binding proteins (CaBPs) consists of dozens of members and contributes to all aspects of the cell's function, from homeostasis to learning and memory. However, the Ca²⁺-binding mechanism is still unclear for most of CaBPs. To identify the Ca²⁺-binding sites of CaBPs, this study presented a computational approach which combined the fragment homology modeling with molecular dynamics simulation. For validation, we performed a two-step strategy as follows: first, the approach is used to identify the Ca²⁺-binding sites of CaBPs, which have the EF-hand Ca²⁺-binding site and the detailed binding mechanism. To accomplish this, eighteen crystal structures of CaBPs with 49 Ca²⁺-binding sites are selected to be analyzed including calmodulin. The computational method identified 43 from 49 Ca²⁺-binding sites. Second, we performed the approach to large-conductance Ca²⁺-activated K⁺ (BK) channels which don't have clear Ca²⁺-binding mechanism. The simulated results are consistent with the experimental data. The computational approach may shed some light on the identification of Ca²⁺-binding sites in CaBPs.