The study reported here explores whether English as a foreign Language (EFL) learners' preferred ways of learning (i.e., learning styles) affect their task performance in computer-mediated communication (CMC). As Ellis (2010) points out, while the increasing use of different sorts of technology is witnessed in language learning contexts, it is worth studying the conditions in which the most second language (L2) production would be accomplished. The participants were 40 advanced-level Iranian EFL learners enrolled at a language institute in Tehran. Learners' individual learning styles were probed by Felder-Soloman (1991) Index of Learning Style (ILS) and they were categorized into 8 groups, within 4 dimensions: Active vs. Reflective, Sensing vs. Intuitive, Visual vs. Verbal, and Sequential vs. Global learners. Then, the participants were given the opinion-gap tasks in 6 consecutive online chat sessions within a 3-week period. The participants' produced language was analyzed at two levels: vocabulary, and grammar. Independent samples t-test were conducted to check if the differences between the groups were significant. The results reveal that the Reflective learners and Visual Learners produced grammatically more complex and lexically denser sentences than the other groups, which suggests that learners' learning styles may affect their task performance in synchronous computer-mediated communication.