The present study investigated the effect of two modes of peer feedback, online and faceto-face, on UTAS Muscat level 4 students' argumentative writing performance and second language writing anxiety. The research design opted for is quasi experimental, where the randomly selected groups (N= 124) were divided to three: control, face-to-face and online. Two main data collection instruments were used. To identify the students' second language writing anxiety levels prior and post the experiment, an instrument titled the Second Language Writing Anxiety Inventory (SLWAI) was adopted from (Cheng, 2004). To determine the students' argumentative writing performance, a posttest and a pretest were designed and employed to fulfill the purpose. The intervention’s duration was a four-week period. Peer feedback was applied on four in-class written essays. In terms of the results, both modes of peer feedback appeared to be effective in improving the students' writing performance and second language writing anxiety. As for the control group, their argumentative writing performance has improved, while their second language anxiety levels remained high. The study's implications indicate that face-to-face peer feedback is more effective to promote a better argumentative writing performance level, while online peer feedback is recommended for reducing second language writing anxiety levels. Future research can be conducted on a bigger scale, a different local context and on different language skills such as reading, listening and speaking.