In order to quantify the contribution of energy storage equipment to system carbon emission reduction and its impact on system flexible peak shaving, a bi-level optimization model of energy storage considering system carbon emission and system regulation risk is established. Firstly, the output carbon emission model of thermal power units is improved and the carbon emission generated by thermal power units is described in detail, then a carbon emission measurement model considering the charge and discharge power of energy storage is proposed. The upper model determines the charging and discharging power of energy storage at each time to minimize system carbon emission, load peak valley difference and energy storage operation cost. Then, a wind power operation risk cost function with adjustable energy storage capacity is established. The lower model aims at minimizing the risk cost of load shedding, wind abandonment and power regulation of thermal power units, and determines the energy storage reserve capacity at each time under the constraint of the energy storage charge and discharge power obtained from the upper layer. In terms of algorithm, immune genetic algorithm is used to solve the upper problem, and the CPLEX solver in MATLAB and point estimation method are used to jointly solve the lower problem. The simulation results show that the model can reduce the peak valley difference of load and reduce the system carbon emission, so as to quantify the contribution of energy storage to emission reduction. By setting the reserve capacity of energy storage, the peak shaving flexibility of the system is improved and the operation risk of the system is reduced. The upper energy storage power optimization model and the lower energy storage reserve capacity optimization model constrain each other, so as to provide reference for low-carbon and flexible operation of the system.