Release rate is one of the important items for the environmental impact assessment caused by radioactive materials in case of an accidental release from the nuclear facilities. In this study, the uncertainty of the estimated release rate is evaluated using Monte Carlo method. Gaussian plume model and linear programming are used for estimating the release rate of a source material. Tracer experiment is performed at the Yeoung-Kwang nuclear site to understand the dispersion characteristics. The optimized release rate was 1.56 times rather than the released source as a result of the linear programming to minimize the sum of square errors between the observed concentrations of the experiment and the calculated ones using Gaussian plume model. In the mean time, 95% confidence interval of the estimated release rate was from 1.41 to 2.53 times compared with the released rate as a result of the Monte Carlo simulation considering input variations of the Gaussian plume model. We confirm that this kind of the uncertainty evaluation for the source rate can support decision making appropriately in case of the radiological emergencies.