Modern power systems are structurally complex, consisting of elements such as buses, transformers and transmission lines, which are interconnected and liable to failures that could result in total system blackout. Early detection and quick identification of the elements that are critical to the security and reliability of such interconnected power networks is essential for smart grid applications. This paper proposes a Coupling Strength Matrix (CSM) method, which is based on Network Structural Characteristics Theory and the Relative Electrical Distance (RED) between nodes in the network. Two indices, Bus Coupling Strength (BCS) and Line Coupling Strength (LCS), are proposed in identifying the critical transmission lines and weak nodes within a power network. The proposed method is demonstrated on the standard IEEE 30-bus system and 32-bus Nigerian network. The results obtained compare well with that obtained using conventional closeness centrality-based method. This new approach could be useful in quick identification of weak nodes and critical lines that could result in voltage collapse or island formation during critical outages, with less computational burden.