In this paper, we study respectively empirical likelihood ratio confidence interval for distribution function of nonparametric single population under balance and unbalance ranked set sampling. It proves that with the same sample size, the empirical likelihood ratio confidence interval is shorter under balance ranked set sampling than simple random sampling method, discusses the sampling allocation of empirical likelihood ratio confidence interval for distribution function under unbalance ranked set sampling, and by using of random simulation, compares the different length of confidence interval under the two different sampling methods to demonstrate the advantage of ranked set sampling.