Accurate energy load prediction has a considerable impact on the economic scheduling and optimal operation of integrated energy system. Radial basis function neural network（RBF-NN）model is introduced to predict the short term electric, gas and heating loads of integrated energy system. Firstly, the Copula theory is used to analyze the correlation of electric, gas and heating loads,and the time series model of electric, gas, heating loads and temperature is established. Then, the structure of RBF-NN network is designed and K-means clustering algorithm is adopted to optimize the hidden layer nodes. Finally, the proposed model is verified by the practical data of an integrated energy system in a park in China.Through the comparison of three cases, it is verified that the method proposed in this paper can effectively consider the coupling relationship among electric, gas and heating loads, and improve the prediction accuracy.