Acid-sensing ion channels (ASICs) are trimeric, proton-gated cation channels found in both the central and peripheral nervous system. ASICs are important to pain sensation in relation to tissue acidification (Wemmie et al. Nat Rev Neurosci 2013). Although known to be activated by a reduction in extracellular pH through binding of protons (Waldmann et al. Nature 1997), there is evidence that ASICs are also modulated by membrane lipids. Arachidonic acid (AA) (Smith et al. J Neurosci 2007) and lysophosphatidylcholine (LPC) (Marra et al. EMBO J 2016) have been shown to activate the ASIC3 subtype without requiring a change in pH. Evidence suggest that lipid activation is likely a direct effect of lipid binding, rather than an effect of membrane distortion (Marra et al. EMBO J 2016). However, this mechanism of activation and possible binding sites are unknown. Structures of chicken ASIC1 have been determined in the resting (Yoder et al. Nature 2018), open (Baconguis et al. Cell 2014), and desensitized (Gonzales et al. Nature 2009) states. Using computational homology modelling and molecular dynamics (MD) simulations, we have simulated models of human ASIC1 and ASIC3 channels in a membrane environment containing AA, LPC and POPC. Exploiting the MARTINI coarse grained model and force field (Marrink et al., J. Phys. Chem. B 2007), we have compared lipid interactions with the two subtypes of ASICs for different functional states. We have identified putative binding sites for AA through analysing lipid contacts and residence times. Our data propose that specific residues, which differ between hASIC1 and hASIC3, are crucial for the AA interaction patterns observed in hASIC3. The identified binding sites are further explored through free energy calculations and atomistic resolution MD simulations.