The excitonic effect in semiconductors is sensitive to dopants. Origins of dopant-induced large variation in the exciton binding energy (E(b)) is not well understood and has never been systematically studied. We choose ZnO as a typical high-E(b) material, which is very promising in low-threshold lasing. To the best of our knowledge, its shortest wavelength electroluminescence lasing was realized by ZnO/BeZnO multiple quantum wells (MQWs). However, this exciting result is shadowed by a controversial E(b) enhancement claimed. In this Letter, we reveal that the claimed E(b) is sensible if we take Be-induced E(b) variation into account. Detailed first-principle investigation of the interaction between dopant atoms and the lattice shows that the enhancement mainly comes from the long-distance perturbation of doped Be atoms rather than the local effect of doping atoms. This is a joint work of experiment and calculation, which from the angle of methology paves the way for understanding and predicting the E(b) variation induced by doping.