Migraines are increasingly common and can seriously affect people’s health. Accurate identification of migraine types is essential for effective treatment. This study addresses the problem of migraine classification using K-means clustering. Our methodology involves preprocessing data by identifying outliers and encoding labels. The elbow method determines the optimal number of clusters, allowing us to categorize migraines based on shared characteristics like symptoms and triggers. We then examine the relationships between migraine subtypes within each cluster using a correlation matrix. The results highlight the potential of unsupervised learning with K-means clustering for accurately predicting migraine types. This method provides valuable insights into the variety of migraine presentations and helpful for the effected people and discover which type of migraine occurs most.