As smart grids continue to develop, accurately calculating line losses in low-voltage active distribution networks is essential to improving energy utilization efficiency and the economic operation of the power grid. This article proposes a deep learning method for line loss calculation in low-voltage active distribution networks. First, by analyzing network characteristics and operational data, the main factors affecting line loss are identified. Second, a deep neural network model is used to automatically learn the complex nonlinear relationship between line loss and these factors from historical operating data as training samples. Finally, testing and verification using real data show that this method notably improves the precision and computational efficiency of line loss analysis. The experimental results indicate that, compared with traditional calculation methods, the proposed method achieves higher accuracy and better generalization.