Temple-Baraitser syndrome is a severe childhood-onset developmental and epileptic encephalopathy for which there are no targeted anti-seizure medications. Our group previously showed that this syndrome is caused by gain-of-function mutations in KCNH1, which encodes the voltage-gated potassium channel KV10.1, and screened arthropod venoms to identify modulators that might restore normal KV10.1 function. The identified peptide Ap1a, derived from the venom of the Ecuadorean tarantula Avicularia purpurea, is the most potent modulator of KV10.1 (IC50 200 nM) described to date and a promising therapeutic candidate for controlling seizures in KCNH1 epilepsy. However, Ap1a exhibits off-target effects on the functionally critical cardiac channel KV11.1 (hERG), which poses significant cardiotoxicity concerns. To elucidate the molecular features governing channel selectivity, we systematically investigated structure–function relationships of Ap1a using alanine-scanning mutagenesis, intercystine loop substitution, and C-terminal modification, and tested the effect of each analogue on channel activation using patch-clamp electrophysiology analysis of mammalian cells expressing KV10.1 or hERG. We found that a spatially contiguous stack of aromatic residues on one surface of the peptide likely represents the primary pharmacophore for binding to KV10.1 and hERG, and that C-terminal amidation is also critical for channel binding. Replacement of a surface-exposed loop on the opposite face of the peptide with a flexible linker enhanced hERG activity without compromising KV10.1 activity, supporting a non-binding role for this region of the peptide. This study has advanced our understanding of Ap1a’s pharmacophore, elucidated sequence motifs that may be common to hERG-targeting peptides, and provided a molecular framework for the rational design of KV10.1-selective analogues with reduced off-target activity on hERG that might be useful for seizure control in KCNH1 epilepsy.