This paper examines the relationship of the policies and practices employed by 3 high school reform models--Early College High Schools, Redesigned High Schools, and High Schools That Work--with student success in college preparatory mathematics courses by the end of the 10th grade. Data on policies and practices collected through a survey of school principals in North Carolina are combined with administrative data on student course-taking and performance. The examined policies include course-taking requirements, rigorous instruction, academic support, personalization, and relevance. Results show that implementation of these policies varies across models and that higher levels of implementation of combinations of these policies are associated with improved outcomes.