Abstract: In language testing and assessment, predictive validity of a test is probably considered as one of the most commonly discussed types of test validity because it is primarily dealt with the question of how accurate a test is as a predictor of academic performance. Therefore, this study examines the predictive validity of the International English Language Testing System (IELTS) as it determines the academic success of students, measured by cumulative grade point average (CGPA). The results indicate that IELTS scores seem not to be satisfactory predictors of academic achievement. It showed that IELTS Reading test score is a better fit for students’ academic success. The study also addresses some recommendations of further language test use.