The access conditions for load aggregators to participate in the peaking auxiliary service are one of the elements that determine the effectiveness of grid peaking, which are difficult to formulate due to the influence of various factors. To this end, an optimal design method for the access power of peaking auxiliary service is proposed, which takes into account the influence of the multiple performance of adjustable loads and the finite rationality of the power grid. Firstly, the performance index system of load aggregators participating in peak load regulation auxiliary services is constructed from four aspects of technology, economy, society and environment; second, considering the limited rationality of the grid, the comprehensive performance of load aggregators participating in grid peaking is evaluated based on the cumulative outlook theory; further, taking into account the seasonal impact on the demand of grid peaking, a new energy and load time-slot optimization method is proposed based on the Wasserstein generative adversarial network and forward scenario. Further, taking into account the influence of seasonal demand on grid peaking, the optimal access power for load aggregators to participate in grid peaking is solved based on Wasserstein generation adversarial network and forward scenario based on the curtailment technique to generate the power scenarios of new energy sources and loads in different time periods with the objective of minimizing the comprehensive cost of grid peaking. Finally, the effectiveness of the proposed model is verified based on an actual example in a region.