<p>This paper presents a novel application of trace clustering techniques to judicial data analysis, addressing the high variability and complexity of judicial workflows, an issue that has received limited attention in prior research. Leveraging real-world data from three Brazilian Small Claims Court units, the proposed method integrates preprocessing, data encoding, and clustering algorithms to segment judicial cases into more homogeneous groups and reveal representative behavioral patterns. These patterns facilitate process simplification and enable more effective analysis. The identified clusters corresponded to distinct types of judicial procedures, and their derived process models were evaluated from both structural and behavioral perspectives. Validation by legal domain experts confirmed the method’s effectiveness, highlighting its potential to expose procedural inefficiencies and enhance comprehension of judicial flows. This study contributes empirical evidence supporting the suitability of trace clustering for analyzing flexible and highly variable processes in the judicial domain. Furthermore, the proposed method offers a generalizable framework applicable to other jurisdictions and domains, providing a scalable solution for simplifying complex event logs and supporting data-driven decision-making in process management.</p>