In particle physics, a glueball (also gluonium, gluon-ball) is a hypothetical composite particle. It consists solely of gluon particles, without valence quarks. Such a state is possible because gluons carry color charge and experience the strong interaction between themselves. Glueballs are extremely difficult to identify in particle accelerators, because they mix with ordinary meson states. In pure gauge theory, glueballs are the only states of the spectrum and some of them are stable.

Theoretical calculations show that glueballs should exist at energy ranges accessible with current collider technology. However, due to the aforementioned difficulty (among others), they have so far not been observed and identified with certainty, although phenomenological calculations have suggested that an experimentally identified glueball candidate, denoted f₀(1710), has properties consistent with those expected of a Standard Model glueball.

The prediction that glueballs exist is an essential prediction of QCD as part of the Standard Model of particle physics that has not yet been unambiguously confirmed experimentally.

Experimental evidence was announced in 2021, by the TOTEM collaboration at the LHC in collaboration with the DØ collaboration at the former Tevatron collider at Fermilab, of odderon (a composite gluonic particle with odd C-parity) exchange. This exchange, associated with a quarkless three-gluon vector glueball, was identified in the comparison of proton–proton and proton–antiproton scattering. In 2024, the X(2370) particle was determined to have mass and spin parity consistent with that of a glueball. However, other exotic particle candidates such as a tetraquark could not be ruled out.

In principle, it is theoretically possible for all properties of glueballs to be calculated exactly and derived directly from the equations and fundamental physical constants of quantum chromodynamics (QCD) without further experimental input. So, the predicted properties of these hypothetical particles can be described in exquisite detail using only Standard Model physics that have wide acceptance in the theoretical physics literature. But, there is considerable uncertainty in the measurement of some of the relevant key physical constants, and the QCD calculations are so difficult that solutions to these equations are almost always numerical approximations (calculated using several very different methods). This can lead to variation in theoretical predictions of glueball properties, like mass and branching ratios in glueball decays.

Theoretical studies of glueballs have focused on glueballs consisting of either two gluons or three gluons, by analogy to mesons and baryons that have two and three quarks respectively. As in the case of mesons and baryons, glueballs would be QCD color charge neutral. The baryon number of a glueball is zero.

Double-gluon glueballs can have total angular momentum J = 0 (which are either scalar or pseudo-scalar) or J = 2 (tensor). Triple-gluon glueballs can have total angular momentum J = 1 (vector boson) or J = 3 (third-order tensor boson). All glueballs have integer total angular momentum that implies that they are bosons rather than fermions.

Glueballs are the only particles predicted by the Standard Model with total angular momentum (J) (sometimes called "intrinsic spin") that could be either 2 or 3 in their ground states, although mesons made of two quarks with J = 0 and J = 1 with similar masses have been observed and excited states of other mesons can have these values of total angular momentum.