Based on introducing the empirical likelihood method, a robust empirical likelihood estimator was presented. The weight function and the bounded score were used to limit the influence of outlier on the estimation equation with the constraint condition. The performance of the proposed estimator was studied through simulation. The simulation results show that the proposed robust empirical likelihood estimator has smaller mean square error, compared to ordinary empirical likelihood estimators. At the same time, the proposed estimator also performs better for heavy tailed data sets in the term of mean square error.