Line loss rate is an important economic and technical index comprehensively reflecting power grid planning, production and management, aiming at the problems of slow calculation speed and large error in current calculation methods, a theoretical line loss rate calculation model combining deep confidence network and deep neural network is proposed. The calculation process is transformed into a multi-feature extraction process, and the model is trained by layer-by-layer greedy method and random small batch gradient descent method. Comparative analysis is conducted through calculation examples and traditional models. The results show that the accuracy and efficiency of the proposed method are improved compared with the traditional line loss rate calculation method, which shows the superiority of the proposed method and has certain practical value.