<p>Under specialized operational scenarios, the deployment of power supply vehicles (PSVs) for energy replenishment of unmanned equipment has emerged as a predominant technological trend. We formally define this optimization challenge as the power supply vehicle routing problem (PSVRP), that entails the strategic deployment of PSVs originating from centralized hubs to dynamically replenish unmanned equipment that have been depleted. Building on this definition, this problem is formulated as a two-stage mixed-integer nonlinear programming model. To solve this model, we adopted a genetic algorithm as the main algorithm framework. We developed crossover operator, mutation operator, and local adjustment operation. These components were then sequentially combined to form a novel evolutionary algorithm, which is referred to as ESGA. Finally, two sets of simulation cases demonstrate that the operators and elite strategy we designed are more competitive in solving the PSVRP. Parameter sensitivity tests indicate that the algorithm we proposed have low sensitivity to their respective parameters. The ablation experiment shows that the crossover operator has a greater impact on ESGA, the mutation operator is beneficial for improving population diversity, and the local adjustment operation can improve the quality of the optimal solution.</p>