We present theory and experiments which describe charge transfer from the X³Σg⁻ and a(1)Δg states of molecular oxygen and atomic and molecular cations. Included in this work are new experimental results for O2(a(1)Δg) and the cations O(+), CO(+), Ar(+), and N₂⁺, and new theory based on complete active space self-consistent field method calculations and an extended Langevin model to calculate rate constants for ground and excited O2 reacting with the atomic ions Ar(+), Kr(+), Xe(+), Cl(+), and Br(+). The T-shaped orientation of the (X - O2)(+) potential surface is used for the calculations, including all the low lying states up to the second singlet state of the oxygen molecule b¹Σ(g)⁺. The calculated rate constants for both O₂(X³Σg⁻) and O2(a(1)Δg) show consistent trends with the experimental results, with a significant dependence of rate constant on charge transfer exothermicity that does not depend strongly on the nature of the cation. The comparisons with theory show that partners with exothermicities of about 1 eV have stronger interactions with O2, leading to larger Langevin radii, and also that more of the electronic states are attractive rather than repulsive, leading to larger rate constants. Rate constants for charge transfer involving O2(a(1)Δg) are similar to those for O₂(X³Σg⁻) for a given exothermicity ignoring the electronic excitation of the O2(a(1)Δg) state. This means (and the electronic structure calculations support) that the ground and excited states of O2 have about the same attractive interactions with ions.