5G network is considered as enabler for Smart Grids (SG). However, it is inefficient and costly to travel over large area and penetrate physical blind area in SG. The significant improvement in this work is proposing 5G LoRa converged network to offer a seamless, large scale, zero wiring, and real-time IoT monitoring strategy. On the basis of this framework, multiple LoRa Ad-Hoc network can coalesce with the 5G SG network via 5G LoRa intelligent gateway. Experiments and tests in real scenarios demenstrate that the proposed 5G LoRa Ad-Hoc networking enables the achievement of communication blind and wide-area SG IoT monitoring. The collection of information is accurate and the system runs stably. By introducing Charnes-Cooper transform and Lagrange Multiplier iteration algorithm, the resource allocation problem is formulated as a mixed integer relaxation optimization, aiming to maximize the sum energy efficiency of multi-hop LoRa. Simulation results unveil that the optimal number of active LoRa nodes and resource allocation in each cluster exists to maximize the area energy efficiency. Analysis also validates the existing ultimate efficiency of our proposed model, which is 6.2 times higher than that of static benchmark algorithms in terms of energy efficiency and QoS requirements.