L2 synonyms (and other related words) that are undifferentiated (i.e., are translated as the same word) in learners’ L1 are among the most difficult lexical items for L2learners to grasp. How to help students learn the appropriate use of these words has been a great challenge for L2 teachers. Drawing on recent research in corpus linguistics and corpus-based language teaching, this paper strives to show that corpus analysis can effectively differentiate synonyms and help L2 students successfully learn the appropriate use of the words in a synonym set. The paper begins with a brief overview of research on synonyms and their learning/use. Then using four sets of English synonyms/related words that are not differentiated in Korean, it discusses, with specific examples of learning activities and exercises, how corpus analysis and its results can be used to help learners differentiate these difficult words and grasp their usage patterns. The paper concludes with a summary of both the benefits and challenges of such a teaching approach and some strategies for dealing with the challenges.