Abstract: In the context of the new power system, timely warnings of the operation situation of the distribution network are an important prerequisite for ensuring the safety and stability of distribution network operation. This paper proposes a distribution network operation situation warning method based on data augmentation in response to the problem of insufficient measurement data for distribution network operation situation warning. Firstly, based on the support vector data description, the initial warning boundary is identified, combined with the fuzzy membership function to enhance the influence of data sample density on boundary recognition and accurately determine the initial warning boundary. Secondly, the data augmentation model of distribution network operation status electrical quantity based on improved deep convolutional generative adversarial networks is established. The loss function introduces deviation of boundary shape, deviation of sample probability distribution, and deviation of situation prediction regression relationship curve to correct the operation status warning boundary of the distribution network. Then, the discrimination basis for the situation warning method is proposed, and the trend of action towards the boundary at the operating point is obtained to achieve distribution network operation situation warning online. Finally, the improved IEEE 33-bus system validation and IEEE 123-bus system validation show that this method can effectively improve the speed and accuracy of distribution network operation situation warnings.