Clustering is a fundamental task in data analysis that involves grouping similar data points to uncover meaningful patterns within datasets. While traditional clustering methods such as K-means and fuzzy C-means are widely used due to their simplicity, they often suffer from limitations such as sensitivity to initial conditions and premature convergence. To address these issues, this study introduces an improved optimization-based approach called Chaotic Levy Flight Siberian Tiger Optimization (CLFSTO). The proposed method enhances the standard Siberian Tiger Optimization algorithm by integrating chaotic logistic maps and Lévy flight strategies, which together improve exploration and exploitation capabilities during the clustering process. CLFSTO is evaluated on ten real-world benchmark datasets with varying dimensionality and complexity. Clustering performance was evaluated using the silhouette score to measure intra-cluster cohesion and inter-cluster separation. On average, CLFSTO achieved 6.69% improvement over STO, 4.62% over CSTO, and other well-known methods of mathematics. Furthermore, Wilcoxon and Friedman statistical tests confirmed that these improvements are statistically significant (p < 0.05). Results demonstrate that CLFSTO consistently outperforms both traditional clustering techniques and several existing metaheuristic algorithms in terms of accuracy and stability, providing a robust and adaptive approach for real-world clustering and data-driven engineering applications.