Molecular imprinting has been extensively studied and applied as a simple technique for creating artificial polymer-based recognition gels for a target molecule. Although this technique is effective when targeting small molecules, attempts to extend it to larger templates, such as proteins, have, for the most part, failed to show similar success. Our group has. developed a simple simulation model to study protein imprinting. In our previous studies, we investigated the structure of the protein-imprinted pore and imprinting factors of various model proteins. Here, we concentrate on imprinting conditions that affect the separation factor, or the ratio between the interaction energies of the template and a competitor protein. We study the effect of size, charge density, and surface charge distribution of the template protein and calculate the separation factor for various polymerization conditions. Our model captures the known effect of increasing imprinting factor (ratio of binding of the protein in an imprinted polymer to that of a nonimprinted polymer) with increasing surface functionality of the polymer but at the cost of reduced selectivity. Most interestingly, we observe that the surface charge distribution of the protein plays an important role in selectivity of the protein-imprinted polymer, suggesting that some proteins may be better candidates for molecularly imprinted polymers than others.