In the era of rapidly evolving communication technologies, terahertz communication systems have emerged as a pivotal solution for breaking through existing communication bottlenecks, owing to their ultra-wide bandwidth, ultra-high data rates, and low latency characteristics. Within communication system architectures, channel coding and decoding technologies-particularly Low-Density Parity-Check (LDPC) codes-play a vital role in enhancing the anti-interference capability and transmission accuracy of communication links. This paper conducts a theoretical analysis of fundamental LDPC coding principles, introduces QC-LDPC (Quasi-Cyclic LDPC) codes, and specifies the (1944, 1620) code structure adopted in this study. The research focuses on investigating relevant decoding algorithms, ultimately selecting the layered normalized min-sum algorithm as the primary decoding approach. Furthermore, the decoding performance of QC-LDPC in millimeter-wave channels has been validated through MATLAB-based simulations of terahertz communication systems.