This study aims to investigate the effects of school-level and student-level variables on Mathematics achievement of Turkish students. The study adopted quantitative research design. The participants were composed of 4498 students in 146 schools. Both schools and students were selected randomly, and they took part in the study on voluntary basis. The data were collected using a school questionnaire, student background questionnaire and Mathematics achievement test of "Trends in International Mathematics and Science Study". The data were analyzed via Hierarchical Linear Modeling. The results showed that 45% of variance between schools, 54.6% of the variance was in schools, 57.33% of school the variance in Mathematics achievement accounted by principals' report on percentages of students coming from economically disadvantaged homes, parents to volunteer for school programs, school resources for mathematics instruction and principals' perception of school climate.