Owing to their extraordinary magnetic properties and low-cost production, iron oxide nanoparticles (IONs) are in the focus of research. In order to better understand interactions of IONs with biomolecules, a tool for the prediction of the propensity of different peptides to interact with IONs is of great value. We present an effective implicit surface model (EISM), which includes several interaction models. Electrostatic interactions, van der Waals interactions, and entropic effects are considered for the theoretical calculations. However, the most important parameter, a surface accessible area force field contribution term, derives directly from experimental results on the interactions of IONs and peptides. Data from binding experiments of ION agglomerates to different peptides immobilized on cellulose membranes have been used to parameterize the model. The work was carried out under defined environmental conditions; hence, effects because of changes, for example structure or solubility by changing the surroundings, are not included. EISM enables researchers to predict the binding of peptides to IONs, which we then verify with further peptide array experiments in an iterative optimization process also presented here. Negatively charged peptides were identified as best binders for IONs in Tris buffer. Furthermore, we investigated the constitution of peptides and how the amount and position of several amino acid side chains affect peptide-binding. The incorporation of glycine leads to higher binding scores compared to the incorporation of cysteine in negatively charged peptides.