Binary Underdeveloped regions are regencies whose areas and communities are less developed compared to other regions on a national scale. In Indonesia, there are 62 underdeveloped regencies scattered across various provinces. This study aims to classify these 62 regencies based on the Indicators for Determining Underdeveloped Regions, which include Gross Regional Domestic Product (GRDP) per capita (X1), Percentage of Non-Food Household Expenditure (X2), Junior High School Participation Rate (X3), Senior High School Participation Rate (X4), Villages with Health Facilities (X5), Villages with Doctors (X6), Villages with Elementary Schools (X7), Villages with Junior High Schools (X8), and Regional Original Revenue (PAD) per capita (X9). The method used in this study is ModelBased Clustering using a multivariate t-distribution approach. This method focuses on a statistical model and is based on the Finite Mixture assumption. In the Finite Mixture framework, the data is assumed to originate from several distributions, and the resulting clusters represent these probability distributions. The study identifies the UUUC model as the best model, producing two optimal clusters with distinct characteristics: Cluster 1 with a low level of regional underdevelopment, and Cluster 2 with a high level of underdevelopment. It is hoped that regencies classified in the highly underdeveloped cluster can be prioritized to achieve equitable development more quickly and effectively