Energy storage system (ESS) can provide services such as peak load shaving and valley load filling, short-term power support for the distribution system with a high percentage of new energy penetration. However, there are few studies on the location and capacity optimization of energy storage considering system reliability. Therefore, this paper first uses analytical methods to model the reliability of traditional equipment and ESS. Then the reliability indicators and calculation methods are proposed based on the charging and discharging efficiency of ESS. Overall optimization considering reliability and cost is the objective function, and genetic algorithm is used to optimize the location and capacity of ESS in the distribution network. Finally, the power distribution system of an industrial park in the south is taken as an example, simulation calculations are performed on Matlab. The reasonable layout of the ESS can reduce the average outage time decrease from 0.32 h/per each outage to 0.24 h/per each outage, which significantly improves the system reliability and verifies the effectiveness, technical and economic efficiency of the proposed model.