<p>Virtual communities within social networks mirror localized characteristics, interpersonal relationships, and individual behavioral patterns, offering insights into network topology and the dynamics of complex systems. However, the inherent diversity of opinions within these networks can pose significant risks to user security and privacy. Consequently, the establishment of secure communication channels is paramount. While existing algorithms have attempted to mitigate these risks, they frequently fall short in accurately identifying trusted communities. This paper addresses this deficiency by proposing a novel methodology for the detection of trusted communities within social media platforms. In the proposed method for community detection, in addition to utilizing new dynamic criteria, including Kullback–Leibler divergence, reputation, common trust, and mutual trust, one of the optimization algorithms is also used. The primary contribution of this work lies in its novel approach to the identification and consolidation of trusted communities, resulting in demonstrably improved efficiency. A comparative analysis on four real-world datasets, against several established algorithms, confirms the effectiveness and efficiency of the proposed method in identifying trusted communities.</p>