Background: Stable headache diagnosis classification is a prerequisite for identification of headache type specific risk factors. Does the stability of a headache diagnosis over time vary between migraine and tension-type headache (TTH)? Are there differences in diagnosis stability between a probable and a definite headache diagnosis? Findings: In a sample of 783 students (ages 12 to 18 years) participating in a headache intervention study in greater Munich, the stability of headache classification according to the International Classification of Headache Disorder - third edition (beta version) (ICHD-3 beta) after a follow-up of 7 months was examined. Differences in stability of probable or definite migraine and probable or definite TTH were assessed. The stability of the headache diagnosis was assessed as predictive value of headache diagnosis with regard to confirmation of the headache type using the same diagnostic instrument 7 months later. Predictive values with 95% confidence intervals (CI) are reported. Of students with initial migraine, a diagnosis of migraine was confirmed in 65.71% of students after 7 months (95%-CI {[}59.40-71.64]). A clear distinction between probable (44.71%, 95%-CI {[}33.91-53.89]) and confirmed diagnosis (76.88% 95%-CI {[}69.56-83.17]) of migraine was observed. For TTH the predictive value was 62.66% (95%-CI {[}57.07-68.01]) overall with a lower stability for probable (46.10%, 95%-CI {[}37.68-54.69]) compared to the confirmed diagnosis (69.71%, 95%-CI {[}23.58-37.67]). Conclusion: While confirmed migraine and confirmed TTH diagnoses seem stable over time, stability of a probable diagnosis for either headache type was lower.