Hydrogen sulfide (H2S) emitted from various sources is a major odorous compound, and non-thermal plasma (NP) has emerged as a promising technique to eliminate H2S. This study was conducted to investigate lab-scale and pilot-scale NP reactors using corona discharge for the removal of H2S, and the effects of relative humidity, applied electrical power on reactor performance and ozone generation were determined. A gas stream containing H2S was injected to the lab-scale NP reactor, and the changes in H2S and ozone concentration were monitored. In the pilotscale NP experiment, the inlet concentration and flow rate were modified to determine the effect of relative humidity and applied power on the NP performance. In the lab-scale NP experiments, H2S removal was found to be the 1st-order reaction in the presence of ozone. On the other hand, when plasma reaction and ozone generation were initiated after H2S was introduced, the H2S oxidation followed the 0th-order kinetics. The ratio of indirect oxidation by ozone to the overall H2S removal was evaluated using two different experimental findings, indicating that approximately 70% of the overall H2S elimination was accounted for by the indirect oxidation. The pilotscale NP experiments showed that H2S introduced to the reactor was completely removed at low flow rates, and approximately 90% of H2S was eliminated at the gas flow rate of 15 m3/min. Furthermore, the elimination capacity of the pilot-scale NP was 3.4 g/m3·min for the removal of H2S at various inlet concentrations. Finally, the experimental results obtained from both the lab-scale and the pilot-scale reactor operations indicated that the H2S mass removal was proportional to the applied electrical power, and average H2S masses removed per unit electrical power were calculated to be 358 and 348 mg-H2S/kW in the lab-scale and the pilot-scale reactors, respectively. To optimize energy efficiency and prevent the generation of excessive ozone, an appropriate operating time of the NP reactor must be determined.