The key to solving the problem of large-scale regional wind and photovoltaic connections to the grid, which pose a threat to the stability and flexibility of the power system, lies in the rational configuration of energy storage capacity. In this paper, a robust model of energy storage was constructed for the clustering photovoltaic-wind joint power generation taking into account the temporal and spatial clustering effect. Added temporal and spatial clustering constraints to the polyhedron uncertainty set describing the wind and solar output to shrink the set boundary and reduce the conservativeness of the model; used mathematical reasoning to transform the uncertainty problem into a deterministic problem, then the relationship between the probability of violation and the confidence level of uncertainty was discussed, and the economics of the system was analyzed based on carbon trading. The model was solved using a double swarm peudo parallelism sparrow search algorithm and differential evolution. The results show that the energy storage configuration scheme obtained by considering the temporal and spatial clustering effect can better balance the economy and robustness of the system; taking carbon trading into account will reduce the cost of the system; uncertainty confidence probabilities are negatively correlated with the probability of violating safety constraints and positively correlated with system economics.