Abstract: This study is aimed at determining the importance of structural descriptors on bioactivity of antifungal peptides by ANN methodology. Due to vital need for novel fungicidal agents, understanding structural features of antifungal peptides and their mechanism of action is necessary. The main AFP actions are interaction with fungal membrane components, formation of membrane channels and pores and eventually fungal cell membrane degradation. Since these mechanisms depend on structural features of the peptides, study of the relationship among these features can give new insights into mode of action of antifungal peptides. In this study, Artificial Neural Network (ANN) is used to determine the key parameters in bioactivity of antifungal peptides. Among a set of 12 molecular descriptor for nearly 60 naturally originated antifungal peptide, α-helical content, hydrophobic interaction and charge positivity introduced as the most important parameters in AFPs activity. This result confirm the barrel-stave pore model (BSPM) as the most probable action mechanism of antifungal peptides.