Nano-grid (NG) clusters (NGCs) have the potential to serve as as an essential configuration for future low-voltage distribution networks. To address the comprehensive energy efficiency (CEE) problem over a longer time horizon, this paper proposes an energy optimization management strategy. First of all, three typical models are developed, namely equivalent circuit model, cluster loss model and energy sharing model. In the modelling, influences of port characteristics and power transmission losses of interconnecting converters are considered to more accurately describe the cluster power flow. Then, an optimal energy scheduling mechanism with high CEE is proposed based on the three established models. The strategy decomposes the complex CEE problem into three interdependent sub-problems, corresponding to the three established models. First, ensuring the bus power quality and battery requirements are satisfied based on the proposed equivalent circuit model. Second, improving cluster self-sufficiency based on cluster energy sharing model. Third, achieving high energy transfer efficiency based on cluster loss model. Extensive simulation results demonstrate that the CEE under the proposed strategy achieves significant performance improvements, with increases of 8% and 32% in energy transfer efficiency and energy self-sufficiency respectively.