Abstract: Taking the typical industrial users of Zhejiang Province as the research object, the potential for participating in the demand response of industrial cluster loads is analyzed, and a demand response model has been established; considering the uncertainty of the aggregate participation in the electrical energy market and the peak adjustment assist service market, including the probability of winning bids, calling probability and shortage possibilities, and use Bayesian neural network to predict, establish a bidding strategy model for aggregates to participate in the joint market, and obtain the purchase of the park in the spot market in each period of time, through the minimum cost of electricity purchase and maximum demand response income, the power purchase volume of the park in the spot market in each period is obtained. The established model provides more accurate and practical decision-making basis for industrial clusters and aggregators, enabling market participants to choose bidding time and capacity more wisely. At the same time, it minimizes the risk of power shortage and maximize total profit, which provides important reference for formulating strategies for industrial cluster load aggregators to participate in the power market.