This study aimed to determine differences in students' argumentation abilities on the subject matter of cell structure and function through a scientific approach in high schools with different accreditation ratings. The research design used is "ex post facto." The sampling technique used was purposive sampling, with a total sample of 111 students consisting of high school students accredited A, namely Seputih State High School 1 Surabaya, high school students accredited B, namely Senior High School Bangun Cipta and high school students accredited C, namely Senior High School Miftahul Ulum. Research data in the form of test results describing argumentation skills were analyzed using the Anova and LSD tests at a significant level of 5%, student questionnaire data were analyzed descriptively in the form of percentages, data from teacher interviews and documentation studies were analyzed descriptively qualitatively using Miles and Huberman's model. The results showed that there was a significant difference in the ability to argue between high school students accredited A and B and C (sig. p<0.05). The argumentative abilities of senior high school students accredited B were not significantly different from senior high school students accredited C (sig. p> 0.05). The average argumentation ability of students from senior high school accredited A is higher than school accredited B and C, but most of the student scores from the senior high schools are still in the "very poor" category. Senior high school students accredited A, B and C are already able to make claims, but the grounds, warrants and backings given are not relevant to the claim. This is due to the application of learning that does not direct students to argue, the application of a scientific approach that is not optimal, and the characteristics of students'.