Land and Building Tax (PBB) is an important source of revenue for local governments. However, taxpayer compliance levels vary across regions, which makes the process of mapping and evaluation more challenging. This study aims to group PBB taxpayers in Bukit Kecil District based on their compliance levels using the K-Medoids algorithm. The research was conducted following the CRISP-DM stages, which include business understanding, data understanding, data preparation, modeling, evaluation, and deployment. The data used consist of PBB taxpayer data from 2018 to 2022 with attributes including Tax Object Number (NOP), sub-district, tax year, and payment status. Before the clustering process, the data underwent preprocessing stages such as cleaning, transformation, and normalization. The results indicate that the K-Medoids method is able to form several groups of taxpayers with different levels of compliance. This grouping provides an overview of taxpayer compliance patterns in the study area. Based on these results, it can be concluded that the K-Medoids algorithm can be used as an analytical tool to support evaluation and strategic planning of PBB management by the Regional Revenue Agency.