While a negative relation between math anxiety (MA) and math performance among adolescents is well-documented, this relation isstill poorly understood among primary-school students. Students with MA (prevalence of 11–17%) experience feelings of tension andanxiety that interfere with their math performance. In the iFeelMath project, we bring new scientific knowledge into this field byinvestigating the developmental and situational dynamics of MA and math performance among primary-school students. Further, to prevent the possible negative consequences of MA, we develop and test the effectiveness of brief classroom interventions to supportstudents with high MA. A sample of 400 fourth grade students are followed over five timepoints until the sixth grade, to investigate the causal ordering of MA and performance. Situational dynamics between MA and task performance will be investigated in a subsampleof 100 students during eight lessons and linked to the overall longitudinal design. Both self-reported situational MA and physiological responses (skin conductance) will be collected when students are working with digital math tasks. Further, three brief interventions (i.e.,mindfulness, math skills and a combined) aimed at reducing MA and improving math performance will be tested with another sample of students (n = 400) using a pre-post-delayed posttest design. The novelty of our combined study designs (incl. longitudinal,situational, and intervention) is expected to provide answers that increase our understanding of when, how, and why MA develops during primary school years. This project is the first to incorporate a situational perspective to advance our understanding on the mechanisms of the MA-performance relation. Inclusion of objective physiological measures, instead of only relying on self-reports, is expected to advance the MA research field. The intervention programs to be developed and tested are a much needed add-on to the MAresearch field, to reveal effective ways to support the primary-school students experiencing high MA. Together, these studies advance our knowledge on which theoretical accounts concerning the MA–math performance link are the most plausible.