Occupancy models are commonly used in ecological surveys to account for species detection issues. This study introduces a new method called empirical likelihood (EL) for estimating the occupancy rate of a species. The researchers found that the EL approach provides more accurate confidence intervals compared to traditional methods, especially when the species detection probability is low. The study also shows that the maximum EL estimator is the same as the maximum likelihood estimator, and the EL ratio statistic follows a chi-square distribution. An expectation-maximization algorithm was developed to facilitate computations.