This study examined mathematics and science performance of students who frequently, sometimes or never spoke Persian at home. The data were obtained from 1914 Iranian fourth-grade students who participated in TIMSS 2003. One-way ANOVA and KruskalWallis statistical tests were used to analyze the data. The results showed mathematics and science performance differs significantly among students who frequently, sometimes or never spoke Persian at home. Students who frequently spoke Persian achieved higher scores in both subjects than those who sometimes or never spoke the language at home. In addition, students who sometimes spoke Persian achieved higher scores in both subjects than those who never spoke it. Furthermore, the results indicated that having access to home educational resources, doing homework on mathematics and science, students’ self-confidence in learning mathematics and science were significantly different among the three groups and the differences were in favour of students who frequently and sometimes spoke Persian respectively.