This paper reports on an empirical study that applied the continuation task to learning the English relative clauses and tested its validity and practicality in this learning. Participants of the study were three groups of university freshmen majoring in English. Group A was required to continue in writing a story text that contains relative clauses whereas Group B continued the same story text having no relative clauses. Group C was asked to read the text, followed by a comprehension check, without having to do the continuation. The study adopted a pretest-posttest-delayed posttest design and lasted three weeks. Results showed that i) participants aligned with the preceding text when performing the continuation, as manifested in significantly more uses of relative clauses when they continued the text containing relative clauses in comparison with their continuation of the text containing no relative clauses; ii) the continuation task produced a better learning effect than the reading-only task; iii) those who did not use relative clauses in their continuations after reading the text containing relative clauses, their learning of the clauses exhibited no significant long-term effect. The findings of the study provide useful insights for teaching and learning of English complex grammatical structures.