Abstract: Due to the large number and strong randomness of EVs (electric vehicles), their participation in the peak adjustment demand response of the grid through V2G aggregators is highly uncertain, which not only affects the operation quality of the grid, but also cannot guarantee the economic benefits of aggregators and users. Therefore, an effective response mechanism is urgently needed to guide EV to reasonably participate in the grid service. In consideration of the profitability requirements and development prospects for V2G aggregation businesses, we propose a multi-temporal optimization scheduling strategy that takes into account the reliability of EV peak demand response.Firstly, the day-ahead pre-scheduling of EV demand response is carried out with the goal of optimal integrity-capacity comprehensive index of EV cluster. Secondly, in order to cope with the adverse impact of intra-day EV default uncertainty on peak demand response, the V2G aggregator uses the intra-day default emergency mechanism to conduct emergency response invitation for the whole range of EVs, and then optimizes EV scheduling with the dynamic evaluation model with the optimal response reliability index. Then, the benefits related to peak regulation demand response are evaluated by V2G aggregators and on EVs sides, including the demand response benefits and default costs of the power grid and aggregators, as well as aggregators and EVs. In addition, the surplus of the default capital pool of V2G aggregators is issued to EV with good credit in the form of credit incentives to encourage EV to keep its promise. Finally, a simulation example is used to verify that the proposed optimal scheduling strategy for peak demand response can effectively improve the reliability of EV cluster response and increase the effectiveness of economic benefits for EV and V2G aggregators.