More and more Photovoltaic-Energy Storage Hybrid Systems (PESHS) are embedded to Distribution Networks (DNs), which has brought new resources and challenges to the optimal dispatches of DNs. The influence mechanism of PESHS on voltage and power loss in DNs is presented in this paper, which reveals the important infulence of active and reactive regulating capability of PESHS on voltage regulation and loss reduction for DNs. Therefore, a Double-objective Extended Reactive Power Optimization (DERPO) model that aims at minimum power loss and voltage violation risk for DNs with embedded PESHS is proposed. Unlike traditional reactive power optimization, the active power output of energy storage equipment and power factor of photovoltaic units are added to control variables in this model, which enriches the essence of reactive power optimization of DNs including power supply. Meanwhile, the genetic algorithm NSGA-II is implemented to perform the optimization and the Pareto front of objective function can be worked out. The best compromise solution is achieved by fuzzy method. The results of contrasting simulation between DERPO and other reactive power optimization methods demonstrate that the proposed DERPO model can realize the coordinated optimization of active and reactive power flow and enhance the voltage safety margin while tapping the potential for reducing loss in active DNs. This work is supported by National High-tech R & D Program of China (863 Program) (No. 2015AA050404) and Science and Technology Project of Guangdong Power Grid Co., Ltd (No. GDKJXM20161607).