<p>A complex network is a symbolic representation of distinct real-world systems where information propagates through nodes. Its goal is to identify communities that represent the network’s structure. However, locating the influential node with the maximal range among various nodes and the ability to disseminate influence to a wide portion of the network is one of the most essential concerns in such a network. Centrality is a traditional metric for understanding the effect of nodes in a network, with numerous variants such as closeness, betweenness, degree centrality, and so on. The centrality metrics either work locally or globally to identify influential nodes. In this study, a proposed algorithm named k-InfNode, based on the characteristics of community structure, captures the dynamics of nodes. k-InfNode uses a random walk and combines local and global properties to figure out which nodes are important in a complex network. It was inspired by the idea of overlapping nodes that show how nodes and communities interact with each other across the network. In the beginning, the fuzzy c-means algorithm finds the overlapping nodes in the network. Next, the algorithm assigns an initial score to each node based on node and community information, and iteratively scores each node using the Random Walk with Restart (RWR) algorithm. Experiments performed using real and artificial networks have shown that the k-InfNode is effective.</p>