We examined the effect of the perceived classroom environment on math self-efficacy and the effect of math self-efficacy on standardized math test performance. Upper elementary school students (N = 1,163) provided self-reports of their perceived math self-efficacy and the degree to which their math classroom environment was mastery oriented, challenging, and caring. Individual student scores on the California Standards Test for Mathematics were also collected. A series of 2-level models revealed that students who perceived their classroom environments as more caring, challenging, and mastery oriented had significantly higher levels of math self-efficacy, and higher levels of math self-efficacy positively predicted math performance. Analysis of the indirect effects of classroom variables on math performance indicated a small significant mediating effect of self-efficacy. Implications for research on self-efficacy and the perceived classroom environment are discussed. (Contains 3 footnotes, 4 tables, and 1 figure.)