Network connectivity, with its significant application value for data transmission and node cooperation, has drawn a great concern in recent years. Facing the heterogeneity and complexity of the IoT system, the connectivity determination between nodes in the network is a big challenge. In view of this, this article proposes a generic and efficient connectivity determination method for IoT applications. This method first characterizes the connectivity parameters of nodes, including the direct connection probabilities between nodes, the degree centrality, and the betweenness centrality of nodes, and based on them, then constructs a node connectivity random graph (NCRG) and splits the NCRG into separate components. Furthermore, it converts the connectivity between nodes located in different components into the connectivity between these components and provides an algorithm to determine their connectivity. Specifically, three testing rules are defined in the algorithm to rank the testing priorities of these components and testing edges between these components. The simulation results show that the proposed method can efficiently achieve high accuracy with less cost.