The current study tested elements of the theoretical model of Portes and Rumbaut (1996), which proposes that parent-child differences in English fluency in immigrant families affect various family processes that, in turn, relate to changes in academic success. The current study of 674 Mexican- origin families provided support for the model in that parent-child fluency in a common language was associated with several dimensions of the parent-child relationship, including communication, role reversal, and conflict. In turn, these family processes predicted child academic performance, school problems, and academic aspirations and expectations. The current findings extend the Portes and Rumbaut (1996) model, however, inasmuch as joint fluency in either English or Spanish was associated with better parent-child relationships. The findings have implications for educational and human service issues involving Mexican Americans and other immigrant groups.