I investigate the effect of teacher subject-specific qualifications on student science achievement using data from TIMSS 2015, a large-scale assessment of student skills. I exploit the availability of student test scores in four different science subjects—biology, chemistry, physics, and earth science—to test whether teachers holding a subject-specific qualification raise student test scores. Using a within-student within-teacher approach, which controls for student and teacher heterogeneity, I find that teacher subject-specific qualification in one subject increases student test scores by 3.5% of a standard deviation in the same subject. The effect is stronger for female students, especially when they are taught by female teachers, for disadvantaged students, and in lower-performing countries. The mediation analysis reveals that 20% of the effect is explained by teachers feeling more confident to teach topics in subjects in which they hold subject-specific qualifications.