<p>One of the most important problems in pattern recognition and data mining is clustering. A popular area of study in computer science and data science is ensemble clustering. Research suggests that a clustering ensemble framework that uses fuzzy clusters may outperform a similar framework that relies on hard clusters. Consequently, our emphasis is on enhancing the proposed method to accommodate fuzzy clusters. This paper presented a novel ensemble clustering algorithm utilizing multiple clusters gained from the improved fuzzy C-means clustering algorithm. To increase diversity in consensus, the C-means fuzzy clustering algorithm is improved. The proposed clustering algorithm to generate Diverse Fuzzy Base Clusters (DFBS) utilized an ensemble of fuzzy clusterings produced by the improved fuzzy C-means clustering method. After that, a weighted graph is created utilizing the ensemble that was obtained in the earlier stage. A final clustering is obtained by partitioning this weighted graph. The suggested DFBS algorithm was compared with other ensemble clustering techniques on 10 different benchmark datasets using various evaluation metrics. Experimental results showed that the suggested DFBS algorithm outperforms other ensemble clustering techniques in most datasets. The suggested approach may efficiently find clusters with varying shapes and improve the performance of the main clustering algorithm.</p>