Analysis of protein dynamics is important to understand its link to biological functions. Classical atomistic methods such as molecular dynamics simulation have been widely used to investigate the protein dynamics in detail, but they are usually limited by the time and length scales that can be simulated. Finite element modeling approach can be employed instead, and has demonstrated its capability to explore the slow, collective motion of proteins efficiently. Here, we present our recent effort to further develop the finite element modeling approach for protein dynamics including model generation, parameter sensitivity, and incorporating solvent damping into the model. Comparison with experimentally measured properties will be also provided..