The purpose of this research is to determine the impact of using multimodal representation modes in the writing practices done by the fifth grade students on their scientific process skills. A combined research method which had both quantitative and qualitative characteristics was used in the research and the groups were chosen as control and experimental groups. 32 5th grade students at an elementary school in the North of Turkey formed the study group. Scientific Process Skills Test and writing activities for learning were used as data collection tools in the study. When the unit was completed, the students were asked to tell their peers about the subjects in the unit with a writing activity in the summary writing type. While the students in the experimental group were obliged to use the representation modes in these writing to learn activities, the students in the control group were asked only to complete the writing activity. In this process, the students performed three writing activities, one preparation and two real practice activities. The criteria determining the content and the features of these writing activities (purpose, subject, interlocutor, writing type and page limit) were standards explained in the instructions distributed to the students. The most basic discrepancy between the groups was the request of using of multimodal representations for the students in the experimental group. The writing activities performed within the scope of the study were scored after they were evaluated in consideration of the criteria determined by the researchers. The Cronbach's alpha credibility coefficient of the scientific process skill test applied as the preliminary test in the beginning of the practice and the post-test at the end of it was determined as 72. The results of the research indicate that using multimodal representations make statistically significant differences on behalf of the experimental group in the scientific process skills of the students.