This study aims to examine whether a maths-unrelated reading intervention can improve children's mathematical performance and to compare the effects of maths-related and maths-unrelated reading interventions. Furthermore, this study examines whether the effects of reading interventions are moderated by individual differences, including gender, pre-test maths performance and general cognitive ability. One hundred thirty-four second-grade Chinese children were recruited into three different groups. In the control group, the children received business as usual teaching, while in the two experimental groups, the children received a maths-related or maths-unrelated reading intervention. The results showed that the maths-unrelated reading intervention could significantly improve children's mathematical performance, and the beneficial effect did not differ from the maths-related intervention. Moreover, the effects of the two reading interventions were not moderated by children's individual differences. These findings support the function hypothesis of language for mathematics.