The proposal of the dual carbon target has made reducing losses and energy conservation a key focus of modern power grid construction. A distribution network line loss estimation method combining convolutional neural networks and improved particle swarm optimization algorithm is proposed to address the problems of low estimation accuracy and poor operational efficiency in existing methods. By improving the particle swarm optimization algorithm to balance individual and global optimal, the optimal weights and thresholds of the network are obtained, which improves the convergence speed and estimation accuracy of convolutional neural network. The feasibility of the proposed line loss rate estimation method is verified through simulation. The results indicate that, compared with conventional methods, the proposed method has higher estimation accuracy and faster operational efficiency, which can provide certain assistance for achieving the dual carbon target.