This study intends to find out the effect of using peer-assessment through instagram for tenth grade students of SMK Walisongo in writing descriptive text. The researcher used quasi-experimental research. Forty seven students were selected as the sample of this research. The sample included two groups, the experimental group and the control group. The experimental group was treated by using peer-assessment through instagram, while the control group was not. The data were gathered through test, either pre-test or post-test. Pre-test was conducted before the treatment that has the purpose to know the basic students’ writing skill in writing descriptive text. Post- test was conducted after the treatment. The goal of post-test was to find out the effect of peer-assessment through instagram towards students’ writing skill. In analyzing the data, the writer used normality test to reveal that the data has the normal distribution, homogeneity test to reveal that the samples has homogenous variant, and independent sample t-test for testing the hypothesis. The result showed the increasing students’ mean score in experimental class, 58,36 for pre-test and 69,80 for post-test. According to the t-test calculation in significance level of 5%, it showed the value of tobservation was higher than ttable (25,05 > 10 ). It meant that the null hypothesis (H0) was rejected and the alternative hypothesis (Ha) was accepted. It concluded that teaching writing by using peer-assessment through instagram is effective and gives the positive result.