Accurate renewable resource (RES) and load prediction play key roles in the power grid planning schemes, eff ective dispatch, and stable operation of power systems. The proportions of wind and solar energy continue to increase, leading to wind and light abandonment. Thus, the absorption of wind and photovoltaic power is particularly important. On the basis of accurately predicting load, wind power and photovoltaic output, the accommodation capacity of wind and photovoltaic power is analyzed. The work contains fi ve parts, as follows: (1) empirical wavelet transform (EWT) is used to decompose wind power and load. At the same time, isolated forest (iForest) and fuzzy C-means clustering (FCM) are used to process photovoltaic data. (2) Low frequency and intermediate frequency components of load are predicted by improved random forest (IRF). High frequency component of load is clustered by improved density-based spatial clustering of applications with noise (IDBSCAN). The processing model is selected on the basis of the characteristics of each class sample. Each component of wind power are predicted by IRF. (3) Photovoltaic power of each category is predicted by IRF. (4) Diff erent components of load and wind power data are added. The photovoltaic power forecast data are synthesized according to the time point. (5) The forecast value of load, wind power, and photovoltaic output of a city are comprehensively evaluated by the summarized prediction level indicators. Three accommodation indicators are used for analyzing the accommodation capacity of wind power and photovoltaic. Results show that the forecasting methods of load, wind power, and photovoltaic power can generate better forecasting results than conventional methods. The analysis results of supplementary prediction level and accommodation indices provide reference for eff ective grid dispatching, sustainable, and healthy energy development.