Community college near-completion students are community college starters who have accumulated a considerable number of credits but left college without any postsecondary educational credential. This quantitative study examined a nationally representative sample and intended to reveal significant predictors of becoming a community college near-completion student. We adopted Bean and Metzner's (1985) framework to focus on characteristics of nontraditional college students and Bahr's (2013) approach to emphasize students' course-taking patterns. We conducted a latent class analysis to explore students' course-taking patterns and examined whether different course-taking patterns would predict the likelihood of being a near-completion student using a logistic regression model. Findings indicated the significant role of course-taking patterns in predicting the likelihood of being a community college near completion student. Community college students who have taken and passed a large number of remedial courses are more likely to leave college without a credential. Additional interaction terms in the regression model further revealed the nuances in terms of the influences of course-taking patterns among various student sub-groups. Implications for practice and future research are discussed.