Abstract: Pulse noise is the most primary factor that affects the performance of power line communication. In order to enhance the performance of polar codes in power line channels, this paper proposes a segmented CA-SCL bit-flipping decoding algorithm and makes simulations in both additive white gaussian noise and power line channels. In constructing the critical set, the algorithm introduces Rate1 sub-block to avoid the complexity of searching the entire set, and segmentation are conducted based on the positions of the bits in the critical set, and each segment flips one bit, achieving multi-bit flipping. Simulation results show that the performance of the SCA-SCL-BF algorithm outperforms that of existing literature significantly when signal-to-noise ratios > 0dB. The algorithm has a great significance in improving power line transmission.