Abstract: Bus load forecasting plays an important role in terms of the safety of power grid dispatching and operating and the accuracy of online analyzing and decision-making. In order to improve the accuracy of bus load forecasting, this paper proposes an ultra-short-term load forecasting method based on multi-type data and hybrid neural network model. Under the background of the big power data, the multi-source data such as the historical bus load data, the date information and the weather information are firstly taken as the input features, and a predictive model is established based on the fusion of the back propagation(BP-ANN) neural network and the convolutional neural network(CNN). Then, the feature vector extraction from the numeric type and category type data of the BP-ANN is concatenated with that of CNN from the image type data. The combined features are fed to a fully connected multilayer Neural Network (NN) to predict the bus load. In this paper, the historical data of active power load at the high voltage side of a 220kV substation in a certain area of China and the corresponding weather information are used for example analysis. The analysis of experimental results shows that the proposed method can effectively utilize the characteristics of multi-source data and model fusion to conduct ultra-short-term bus load prediction, and it has higher load prediction accuracy compared to the single model prediction of either BP-ANN or CNN.