This article presents a comprehensive approach for DNA methylation-based classification of central nervous system (CNS) tumors across all entities and age groups. The authors developed a machine learning algorithm, specifically the Random Forest (RF) classifier, to accurately classify CNS tumors based on DNA methylation profiles. They established a reference cohort of 2,801 samples from 91 histopathological entities and variants, using genome-wide DNA methylation profiles. The RF classifier was trained and validated on this cohort, achieving high discrimination power with an estimated error rate of 4.89% for raw scores and 4.28% for calibrated scores. The classifier was then applied to a prospective cohort of 1,155 diagnostic CNS tumors, leading to a 12% change in diagnosis in up to 12% of cases. The method was also evaluated in external centers, where it helped establish new diagnoses in 12% of cases. The authors developed a free online classifier tool (www.molecularneuropathology.org) to facilitate the application of this method in routine diagnostics. The study highlights the potential of DNA methylation-based classification to reduce inter-observer variability in CNS tumor diagnostics and improve clinical decision-making.This article presents a comprehensive approach for DNA methylation-based classification of central nervous system (CNS) tumors across all entities and age groups. The authors developed a machine learning algorithm, specifically the Random Forest (RF) classifier, to accurately classify CNS tumors based on DNA methylation profiles. They established a reference cohort of 2,801 samples from 91 histopathological entities and variants, using genome-wide DNA methylation profiles. The RF classifier was trained and validated on this cohort, achieving high discrimination power with an estimated error rate of 4.89% for raw scores and 4.28% for calibrated scores. The classifier was then applied to a prospective cohort of 1,155 diagnostic CNS tumors, leading to a 12% change in diagnosis in up to 12% of cases. The method was also evaluated in external centers, where it helped establish new diagnoses in 12% of cases. The authors developed a free online classifier tool (www.molecularneuropathology.org) to facilitate the application of this method in routine diagnostics. The study highlights the potential of DNA methylation-based classification to reduce inter-observer variability in CNS tumor diagnostics and improve clinical decision-making.