The purpose of this study was to investigate the use of English language proficiency and academic reading assessment scores to predict the future academic success of English learner (EL) students. Data from two cohorts of middle-school ELs were used to evaluate three prediction models. One cohort of students was used to develop the prediction equations; the other cohort of students was used to evaluate the accuracy of the prediction models. Classification accuracy at the total group level was higher than 70% for all models. However, classification accuracy was highly dependent upon where the distribution of scores for a particular group was in relation to the cut score. Predication accuracy in terms of the difference between predicted and actual scores was also used to evaluate prediction models. Prediction model accuracy appeared to depend on students' level of English language proficiency. A discussion of how these results could serve a variety of educational and research purposes is included.