This paper assesses the relationship between courses taken in high school and college major choice. Using High School and Beyond survey data, I study the empirical relationship between college performance and different types of courses taken during high school. I find that students sort into college majors according to subjects in which they acquired more skills in high school. However, I find a U-shaped relationship between the diversification of high school courses a student takes and their college performance. The underlying relation linking high school to college is assessed by estimating a structural model of high school human capital acquisition and college major choice. Policy experiments suggest that taking an additional quantitative course in high school increases the probability that a college student chooses a science, technology, engineering, or math major by four percentage points.