Abstract: With the increase of renewable energy penetration, the net load fluctuation becomes larger. Therefore, a flexible ramping optimal scheduling method of integrated energy systems is proposed in this paper for the uncertain change of net load, combining the correlation of prediction error and conditional value at risk (CVaR). First, the initial power of scenic load is predicted by a neural network model, and the joint probability distribution of prediction errors of multivariate random variables is constructed by using the method of the C-vine Copula function for the prediction errors, and a flexible ramping product design method is proposed accordingly. Then, considering the conditional value at risk of scenic abandonment and load cutting brought by net load uncertainty, a flexible ramping product risk value is constructed. Later, an optimal scheduling model of integrated energy system considering the value at risk of flexible ramp climbing product is constructed considering the net load uncertainty. Simulation results show that optimization considering the correlation of load forecasting errors has improved the overall economic efficiency of the system by 1.22%. Both load shedding and curtailed wind power have decreased by 17.01%, validating the effectiveness of the proposed method. The CVaR values at different confidence levels can provide a certain risk reference for the scheduling of integrated energy systems.