Rationally designing proteins that outperform natural enzymes and de novo design of enzymes remain grand challenges after a decades-long effort. Achieving this goal would be a significant test for our understanding of enzyme catalysis and take us beyond the blackbox of directed evolution. In the course of studying liver alcohol dehydrogenase, we found that changing serine to threonine (S48T), the hydrogen bond donor at the active site, or replacing Zn2+ with Co2+, the catalytic metal site, both resulted in higher reaction rates of hydride transfer than that of the wild type. Using vibrational Stark effect spectroscopy, an experimental method enabling us to measure the electric field at the enzyme's active site, we found that both Co2+ and S48T mutations increase the active-site electric field, which shows a linear correlation with the reaction's free energy barrier. Given the correlation established by those single-point mutants, we designed double-point mutants and made quantitative predictions for both their field and activity—those predictions were found in close agreement with the experimental results. In particular, the S48T-Co2+ mutant shows a 50-fold increase in the hydride transfer rate compared to the wild type. We therefore demonstrate that the electrostatic interactions not only drive the catalysis of natural enzymes, but also can be leveraged for improving a natural enzyme. Furthermore, the observation that both serine and Zn2+ mutants follow the same linear correlation indicates that, despite their distinct chemical identities, the catalytic effects of hydrogen bonds and metal ions have a common electrostatic nature and can be quantified in a unifying manner. Taken together, we demonstrate electrostatic catalysis as a physical (and thus rational), quantitative, and predictive framework that unifies different intermolecular forces, for designing enzymes beyond the natural ones.