The article introduces a new algorithm for reducing attributes in data analysis. This algorithm is based on rough set theory and uses a heuristic approach with evaluation index information. By defining distinguishable relations and matrices, the algorithm efficiently identifies and eliminates unnecessary attributes. Theoretical analysis and experimental results confirm the algorithm's accuracy and effectiveness in simplifying complex data sets.