The primary purpose of this study was to examine some variables to predict at-risk learners who are more likely to dropout from a cyber university based on variables related to dropout of cyber university learners. This study collected data from 312 undergraduate students who took ‘adult learning and counselling' course offered at a cyber university in 2008 and 2009. Frequency analysis, factor analysis, and logistic analysis were used to analyze the data. The finding of the study shows that only three variables of ‘the total learning time', ‘curriculum and learning contents', and ‘the number of login' are statistically significant to identify at-risk learners. Some implications are also suggested to lead at-risk learners to succeed in completing a course.