To study the molecular mechanisms of local anesthesia, locations of local anesthetic dibucaine in model membranes and the interactions of dibucaine with a Na+ channel inactivation gate peptide have been studied by 2H- and 1H-NMR spectroscopies. The 2H-NMR spectra of dibucaine-d9 and dibucaine-d1, which are deuterated at the butoxy group and at the 3 position in its quinoline ring, respectively, have been observed in multilamellar dispersions of the lipid mixture composed of phosphatidylcholine, phosphatidylserine, and phosphatidylethanolamine. 2H-NMR spectra of deuterated palmitic acids incorporated, as a probe, into the lipid mixture containing cholesterol have also been observed. An order parameter, SCD, for each carbon segment was calculated from the observed quadrupole splittings. Combining these results, we concluded that first, the butoxy group of dibucaine is penetrating between the acyl chains of lipids in the model membranes, and second, the quinoline ring of dibucaine is located at the polar region of lipids but not at the hydrophobic acyl chain moiety. These results mean that dibucaine is situated in a favorable position that permits it to interact with a cluster of hydrophobic amino acids (Ile-Phe-Met) within the intracellular linker between domains III and IV of Na+ channel protein, which functions as an inactivation gate. To confirm whether the dibucaine molecule at the surface region of lipids can really interact with the hydrophobic amino acids, we synthesized a model peptide that includes the hydrophobic amino acids (Ac-GGQDIFMTEEQK-OH, MP-1), the amino acid sequence of which corresponds to the linker part of rat brain type IIA Na+ channel, and the one in which Phe has been substituted by Gln (MP-2), and measured 1H-NMR spectra in both phosphate buffer and phosphatidylserine liposomes. It was found that the quinoline ring of dibucaine can interact with the aromatic ring of Phe by stacking of the rings; moreover, the interaction can be reinforced by the presence of lipids. In conclusion, we wish to propose that local anesthesia originates from the pi-stacking interaction between aromatic rings of an anesthetic molecule located at the polar headgroup region of the so-called boundary lipids and of the Phe in the intracellular linker between domains III and IV of the Na+ channel protein, prolonging the inactivated state and consequently making it impossible to proceed to the resting state.