The author examined the simultaneous influence of Japanese middle school student and school socioeconomic status (SES) on student math achievement with two-level multilevel analysis models by utilizing the Trends in International Mathematics and Science Study (TIMSS) Japan data sets. The theoretical framework used in this study was Bronfenbrenner's bioecological theory (Bronfenbrenner and Morris 1998). The data sets contained 4,856 students from 146 public and private middle schools. The results indicated that at the student level, different aspects of student SES (i.e., number of books, the possession of computers, paternal, and maternal educational achievements) were positively related to Japanese student math achievement. At the school level, two aspects of school SES (i.e., less populated schools and economically disadvantaged schools) were negatively related to Japanese student math achievement. None of the cross-level interactions were significant, but the random effect for the computer slope was significant. Although this study found both student and school SES effects on student achievement, the proportional reduction of prediction error explained by both student and school SES were was small, meaning the residual variances at student and school levels did not capture the majority of variance explained by math achievement. The implications of theoretical framework and educational policy are discussed.