The aim of this study is comparing math anxiety of secondary school female students in groups (Science and Mathematical Physics) Public Schools, district 2, city of Sari. The purpose of the research is applied research, it is a development branch, and in terms of the nature and method, it is a causal-comparative research. The statistical population in this study consisted of all secondary school female students in groups (Science and Mathematical Physics) Public Schools, district 2, city of Sari that is the equivalent of 1,900 people according to the latest figures cast of Education in Sari city. According to the population size and referring to the Morgan's sample size table, the sample size was considered 320 people. Simple random sampling method was used to select research samples. Data collection methods included library and observation. Data collection tool in the survey is questionnaire. Pelick and Parker (1982) math anxiety questionnaire that their reliability was calculated 0.98 by using Cronbach's alpha. Research Hypothesis test and data derived from questionnaires were analyzed by using SPSS software and descriptive and inferential statistics. The tests used in the research included Kolmogorov-Smirnov test and Pearson correlation and T of two independent groups. The results show a difference between math anxiety among students of mathematics and science. According to the average score of students' math anxiety, we find out students' math anxiety in science is more than math. There is an inverse relationship between math anxiety and students' academic achievement. Students, who have high math anxiety, have lower academic achievement.