Community detection in a social network has become a very important step in the recognition of the structure and dynamics of networks in different areas. At present, overlapping community detection methods have become an important and challenging research field, as they have introduced the prospect of vertex connections in more than one community. As a result, many clustering methods have been proposed to discover communities in social networks. Conversely, conventional node clustering and comparatively novel presented link clustering techniques have some disadvantages during overlapping community detection. Existing agglomerative node clustering is insufficient to capture the persistent overlaps owing to the low throughput and unclear description of communities. Therefore, in this paper, a novel hybrid particle-based agglomerative hierarchical (HPAHC) clustering technique is presented for overlapping communities discovered in complex social networks. In the proposed hybrid clustering method, the node’s similar information or behavior and geometric structure (local and global structure) information is extracted from the link topology in social networks. Subsequently, the particle-based agglomerative hierarchical clustering method utilizes the extracted source information for community detection. In social networks, a significant similarity measure can be designed for every pair of objects by considering both the node behavior and geometric structural features. Thus, nodes with high similarity are preserved in a similar community, and our proposed technique will detect communities effectively whose boundaries cannot be simply divided.