Assumptions about long-term trends in international migration are an increasingly important component of the demographic projection module. Yet most official immigration projections both in the United States and abroad rely on ad-hoc assumptions based on little theory and virtually no definable methodology. The purpose of this paper is twofold: to assess where projection-making agencies stand in their practice of immigration projection and to explore how theoretical insights about immigration may help them improve their practice. The first section describes the current projection methods of leading national and international agencies, from the U.S. Census Bureau and Social Security Administration to the United Nations and the World Bank. The second section scans the wide and varied array of “theoretical frameworks” that attempt to explain international migration flows. The paper identifies six important ones: the policy, the neoclassical, the world systems, the new economics, the social networks, and the dual labor market frameworks. We conclude that much progress might be achievable if the explanatory richness of immigration theory could somehow be consolidated and integrated into a useable projection method. The third section briefly outlines some first steps to start harnessing theory and improving practice.