The docking theory of olfaction proposes that the smell of an odorant molecule is due to a range of weak non-covalent interactions between the odorant [a ligand] and one or more G protein-coupled odorant receptors (found in the nasal epithelium). These include intermolecular forces, such as dipole-dipole and Van der Waals interactions, as well as hydrogen bonding. More specific proposed interactions include metal-ion, ion-ion, cation-pi and pi-stacking. Interactions can be influenced by the hydrophobic effect. Conformational changes can also have a significant impact on interactions with receptors, as ligands have been shown to interact with ligands without being in their conformation of lowest energy.

While this theory of odorant recognition has previously been described as the shape theory of olfaction, which primarily considers molecular shape and size, this earlier model is oversimplified, since two odorants may have similar shapes and sizes but are subject to different intermolecular forces and therefore activate different combinations of odorant receptors, allowing them to be distinguished as different smells by the brain. Other names for the model, such as "lock and key" and "hand in glove", are also misnomers: there are only 396 unique olfactory receptors and too many distinguishable smells for a one-to-one correlation between an odorant and a receptor.

In a seminal paper published in 2023 in Nature which is consistent with the above description of the docking theory, Billesbølle and coworkers use cryo-electron microscopy to determine for the first time the structure of a human OR activated by an odorant, namely OR51E2 activated by propionate. The authors indicate that "propionate binds in a small cavity in OR51E2 that is completely occluded from the external solvent. It binds through two types of contact — specific ionic and hydrogen bonds, and non-specific hydrophobic contacts." Because of the specific shape of the binding pocket, OR51E2 is said to be specific for propionate and "does not bind to fatty acids with longer carbon chains."

The docking theory of olfaction previously relied on the known properties of other G protein-coupled receptors that have been crystallized, as well as structural predictions given the known primary structure, to produce a likely olfactory receptor model. Though olfactory receptors are similar to other G protein-coupled receptors, there are notable differences in the primary structure that make exact comparisons unfeasible. Because of this, predicted olfactory receptor structures have been aided by the development of new structure-predicting software. From this data, simpler odorant-receptor binding models have been developed into more nuanced ideas which consider the distortion of flexible molecules so as to form optimal interactions with binding partners. These modifications help the model to conform better to what is known of the molecular docking of non-olfactory G-protein coupled receptors.

In 1949, R.W. Moncrieff published an article in American Perfumer called "What is odor: a new theory," which used Linus Pauling's notion of shape-based molecular interactions to propose a shape-based theory of odor. This superseded the older vibration theory of olfaction, and, renamed the docking theory of olfaction to more accurately reflect a range of non-covalent interactions in addition to shape, remains the mainstream theory, in both commercial fragrance chemistry and academic molecular biology. Three years after Moncrieff proposed the theory, John Amoore speculated further that the over ten thousand smells distinguishable by the human olfaction system resulted from the combination of seven basic primary odors correlating to odor receptors for each, much as the spectrum of perceived colors in visible light is generated by the activation of three primary color receptors. Amoore's seven primary odors included sweaty, spermous, fishy, malty, urinous and musky. His most convincing work was done on the camphoraceous odor, for which he posited a hemispherical socket in which spherical molecules, such as camphor, cyclooctane, and naphthalene could bind.

When Linda Buck and Richard Axel published their Nobel Prize winning research on the olfactory receptors in 1991, they identified in mice 1,000 G-protein-coupled receptors used for olfaction. Since all types of G-protein receptors currently known are activated through binding (docking) of molecules with highly specific conformations (shapes) and non-covalent interactions, it is assumed that olfactory receptors operate in a similar fashion. Further research on human olfaction systems identified 347 olfactory receptors.

A recent version of the previously named shape theory, also known as odotope theory or Weak Shape Theory, holds that a combination of activated receptors is responsible for any one smell, as opposed to the older model of one receptor, one shape, one smell. Receptors in the odotope model recognize only small structural features on each molecule, and the brain is responsible for processing the combined signal into an interpreted smell. Much current work on the docking theory focuses on neural processing, rather than the specific interaction between odorant and receptor that generates the original signal.

The 2023 cryo-electron microscopy structural study of the binding of propionate to human olfactory receptor OR51E2 published in Nature is fully consistent with the docking theory of olfaction for the particular odorant and receptor involved.