Recent studies have demonstrated that telecommunication networks can be well modeled as complex ones, instead of the classic approach based on graph theory. The study of complex networks is a young and active area of scientific research inspired largely by the empirical study of real-world networks such as computers and social networks. In this paper, a new approach for modeling Wireless Sensor Networks (WSNs) is proposed, with the purpose of ensuring an improvement in terms of clustering coefficient and average network diameter. With the proposed hybrid idea, the major part of nodes will not be necessarily adjacents but, for each pair of them, there will exists a relatively short path for a stable connection. The effectiveness of the proposed idea has been validated through a deep campaign of simulations.