The effectiveness of using peer feedback to improve elementary students' writing through blogging was investigated. The experiment was conducted for 2 hours per week over a period of 15 weeks. An online platform of blogging, uSchool, was provided to the participants for posting their written works. A series of quantitative and qualitative techniques were employed to analyze the data, including interrater reliability, content analysis, t-test, analysis of covariance, observations and interviews. A total of 33 fifth graders were selected as participants and they were assumed to possess Internet skills and the ability to input Chinese characters. The frequencies of peer feedback on the articles were categorized into low, medium and high frequency groups for analysis. The statistical results reached a significant level, indicating that the posttest performance was higher than that of the pretest. It showed that the use of peer feedback through the blogging model is effective for improving the quality of writing.