Pursuant to the criterion of fluency, two types of mathematical achievement tests were used in the present study: simple subtraction (to measure mathematical fluency) and number series completion (to serve as a nonfluency mathematics test). A cohort of 223 d/Deaf and hard of hearing (d/Dhh) students in grades 3-9 in special education schools took a series of cognitive and mathematical tests. After outlying data were considered, the sample was reduced to 198 students; the findings were consistent with expectations: The numerical magnitude processing did not add significantly to the prediction of mathematical reasoning (nonfluency mathematics) but did make a significant contribution to the prediction of arithmetic computation (fluency mathematics) after demographic variables and general cognitive processing were controlled for. The findings suggest that the effect of numerical magnitude processing on d/Dhh children's mathematical performance can be influenced by mathematical fluency.