This study is conducted to determine the effectiveness of teaching grammar-in-context to minimize students' grammatical errors in writing. The design of the study was a quasi-experimental with a non-randomized pretest-posttest control group. The samples of the study were taken from the population of the tenth-grade students. The control group was taught by conventional grammar which was separately given with writing skills. Likewise, the experimental one was treated by teaching grammar-in-context. The results of the study showed that the mean score in the post-test was higher than that in the pretest; and the mean score of experimental group increased 16.20 point after the treatment. This result indicated that teaching grammar in-context is considered to be effective in minimizing students' grammatical errors in writing.