The purpose of this study was to investigate whether self-assessment as a metacognitive strategy has any obvious effect on improving grammatical knowledge of Iranian EFL students. To achieve the purpose, 60 male junior high school students with the age range of 13-16 participated in this study. A Nelson test was administered to homogenize the two groups regarding language proficiency and outliers were discarded. The researcher had one experimental and one control group. To determine students' level of grammatical knowledge a pre-test developed by Farhady (2000) was administered. The result showed that there was no significant difference between the two groups. After the pre-test the treatment began. The control group received no extra instruction and underwent traditional assessment. Whereas the experimental group went through the process of self-assessment whereby they checked their own activities and found their own strengths and weaknesses. Self-assessment was practiced for six sessions during three months. A session after the treatment was over; a post-test developed by Farhady (2000) was administered. The results were compared through an independent t-test. The value of t was 3.417 with a significance of .001 depicting a significant difference between the two groups in the study. The outcome showed that self-assessment had a positive effect on improving students' grammatical knowledge. The findings of this study will be an incentive for syllabus designers as well as teachers and students to allocate specific time to self-assessment. Teachers can use self-assessment along with other activities to improve grammatical knowledge.