In view of typical daily load curve, a method of optimal placement and capacity of distributed energy storage in distribution system based on the sensitivity variance of network loss is proposed. In the method, the charging and discharging operation state of distributed energy storage are taken into account. The priority of all nodes in the distribution network to locate distributed energy storage based on the sensitivity variance of network loss is determined. Taking the distribution network loss and node voltage fluctuation as the objective function, the modified particle swarm algorithm is used to optimize the charging and discharging power of distributed energy storage. And the optimal capacity of distributed energy storage is determined. Finally, the simulation results show that the optimal method can effectively maximize the balance of power on the ground, optimize the network loss of the distribution network and lower the voltage fluctuation of the nodes, and achieve the optimal configuration of distributed energy storage in the distribution network. This work is supported by Science and Technology Project of State Grid Corporation of China (No. KJGW2018-014) and Fundamental Research Funds for the Central Universities (No. 2018QN075).