Supramolecular polymers are a subset of polymers where the monomeric units are connected by reversible and highly directional secondary interactions–that is, non-covalent bonds. These non-covalent interactions include van der Waals interactions, hydrogen bonding, Coulomb or ionic interactions, π-π stacking, metal coordination, halogen bonding, chalcogen bonding, and host–guest interaction. Their behavior can be described by the theories of polymer physics in dilute and concentrated solution, as well as in the bulk.

Additionally, some supramolecular polymers have distinctive characteristics, such as the ability to self-heal. Covalent polymers can be difficult to recycle, but supramolecular polymers may address this problem.

The preamble of the field of supramolecular polymers can be considered dye-aggregates and host-guest complexes. In early 19th century, it was noticed that dyes aggregate via "a special kind of polymerization". In 1988, Takuzo Aida, a Japanese polymer chemist, reported the concept of cofacial assembly wherein the amphiphilic porphyrin monomers are connected via van der Waals interaction forming one-dimensional architectures in solution, which can be considered as a prototype of supramolecular polymers. Soon thereafter, one-dimensional aggregates were described based on hydrogen bonding interaction in the crystalline state. With a different strategyusing hydrogen bonds, Jean M. J. Fréchet showed in 1989 that mesogenic molecules with carboxylic acid and pyridyl motifs, upon mixing in bulk, heterotropically dimerize to form a stable liquid crystalline structure. In 1990, Jean-Marie Lehn showed that this strategy can be expanded to form a new category of polymers, which he called "liquid crystalline supramolecular polymer" using complementary triple hydrogen bonding motifs in bulk. In 1993, M. Reza Ghadiri reported a nanotubular supramolecular polymer where a b-sheet-forming macrocyclic peptide monomer assembled together via multiple hydrogen bonding between adjacent macrocycles. In 1994, Anselm. C. Griffin showed an amorphous supramolecular material using a single hydrogen bond between a homotropic molecules having carboxylic acid and pyridine termini. The idea to make mechanically strong polymeric materials by 1D supramolecular association of small molecules requires a high association constant between the repeating building blocks. In 1997, E.W. "Bert" Meijer reported a telechelic monomer with ureidopyrimidinone termini as a "self-complementary" quadruple hydrogen bonding motif and demonstrated that the resulting supramolecular polymer in chloroform shows a temperature-dependent viscoelastic property in solution. This is the first demonstration that supramolecular polymers, when sufficiently mechanically robust, are physically entangled in solution.

Monomers undergoing supramolecular polymerization are considered to be in equilibrium with the growing polymers, and thermodynamic factors therefore dominate the system. However, when the constituent monomers are connected via strong and multivalent interactions, a "metastable" kinetic state can dominate the polymerization. An externally supplied energy, in the form of heat in most cases, can transform the "metastable" state into a thermodynamically stable polymer. A clear understanding of multiple pathways exist in supramolecular polymerization is still under debate, however, the concept of "pathway complexity", introduced by E.W. "Bert" Meijer, shed a light on the kinetic behavior of supramolecular polymerization. Thereafter, many dedicated scientists are expanding the scope of "pathway complexity" because it can produce a variety of interesting assembled structures from the same monomeric units. Along this line of kinetically controlled processes, supramolecular polymers having "stimuli-responsive" and "thermally bisignate" characteristics is also possible.

In conventional covalent polymerization, two models based on step-growth and chain-growth mechanisms are operative. Nowadays, a similar subdivision is acceptable for supramolecular polymerization; isodesmic also known as equal-K model (step-growth mechanism) and cooperative or nucleation-elongation model (chain-growth mechanism). A third category is seeded supramolecular polymerization, which can be considered as a special case of chain-growth mechanism.

Supramolecular equivalent of step-growth mechanism is commonly known as isodesmic or equal-K model (K represents the total binding interaction between two neighboring monomers). In isodesmic supramolecular polymerization, no critical temperature or concentration of monomers is required for the polymerization to occur and the association constant between polymer and monomer is independent of the polymer chain length. Instead, the length of the supramolecular polymer chains rises as the concentration of monomers in the solution increases, or as the temperature decreases. In conventional polycondensation, the association constant is usually large that leads to a high degree of polymerization; however, a byproduct is observed. In isodesmic supramolecular polymerization, due to non-covalent bonding, the association between monomeric units is weak, and the degree of polymerization strongly depends on the strength of interaction, i.e. multivalent interaction between monomeric units. For instance, supramolecular polymers consisting of bifunctional monomers having single hydrogen bonding donor/acceptor at their termini usually end up with low degree of polymerization, however those with quadrupole hydrogen bonding, as in the case of ureidopyrimidinone motifs, result in a high degree of polymerization. In ureidopyrimidinone-based supramolecular polymer, the experimentally observed molecular weight at semi-dilute concentrations is in the order of 10⁶ Dalton and the molecular weight of the polymer can be controlled by adding mono-functional chain-cappers.

Conventional chain-growth polymerization involves at least two phases; initiation and propagation, while and in some cases termination and chain transfer phases also occur. Chain-growth supramolecular polymerization in a broad sense involves two distinct phases; a less favored nucleation and a favored propagation. In this mechanism, after the formation of a nucleus of a certain size, the association constant is increased, and further monomer addition becomes more favored, at which point the polymer growth is initiated. Long polymer chains will form only above a minimum concentration of monomer and below a certain temperature. However, to realize a covalent analogue of chain-growth supramolecular polymerization, a challenging prerequisite is the design of appropriate monomers that can polymerize only by the action of initiators. Recently one example of chain-growth supramolecular polymerization with "living" characteristics is demonstrated. In this case, a bowl-shaped monomer with amide-appended side chains form a kinetically favored intramolecular hydrogen bonding network and does not spontaneously undergo supramolecular polymerization at ambient temperatures. However, an N-methylated version of the monomer serves as an initiator by opening the intramolecular hydrogen bonding network for the supramolecular polymerization, just like ring-opening covalent polymerization. The chain end in this case remains active for further extension of supramolecular polymer and hence chain-growth mechanism allows for the precise control of supramolecular polymer materials.

This is a special category of chain-growth supramolecular polymerization, where the monomer nucleates only in an early stage of polymerization to generate "seeds" and becomes active for polymer chain elongation upon further addition of a new batch of monomer. A secondary nucleation is suppressed in most of the case and thus possible to realize a narrow polydispersity of the resulting supramolecular polymer. In 2007, Ian Manners and Mitchell A. Winnik introduced this concept using a polyferrocenyldimethylsilane–polyisoprene diblock copolymer as the monomer, which assembles into cylindrical micelles. When a fresh feed of the monomer is added to the micellar "seeds" obtained by sonication, the polymerization starts in a living polymerization manner. They named this method as crystallization-driven self-assembly (CDSA) and is applicable to construct micron-scale supramolecular anisotropic structures in 1D–3D. A conceptually different seeded supramolecular polymerization was shown by Kazunori Sugiyasu in a porphyrin-based monomer bearing amide-appended long alkyl chains. At low temperature, this monomer preferentially forms spherical J-aggregates while fibrous H-aggregates at higher temperature. By adding a sonicated mixture of the J-aggregates ("seeds") into a concentrated solution of the J-aggregate particles, long fibers can be prepared via living seeded supramolecular polymerization. Frank Würthner achieved similar seeded supramolecular polymerization of amide functionalized perylene bisimide as monomer. Importantly, the seeded supramolecular polymerization is also applicable to prepare supramolecular block copolymers.