LDPC codes 在消息理論上越來越獲得重視，然而一個主要的問題就是面對這些碼在傳輸系統裡是高解碼複雜度導致需要較長的解碼時間，像這樣較長的解碼時間在許多應用上是不被接受的。利用訊息傳遞解碼演算法，我們發現在解碼方面有很好的效能，但因為其複雜度偏高所花的解碼時間也比較長，所以，很多研究都在討論如何改善複雜度，又不會使效能嚴重的衰減。相對於硬式解碼演算法在效能方面不像軟式解碼演算法這麼出色，不過，硬式解碼演算法也有優點，像是複雜度低、易於硬體實現且解碼時間較快。在這篇論文，主要是討論藉由混合硬式與軟式解碼演算法使其在效能與解法時間上取得一個平衡，讓效能可以維持一定的水準，且使用較少的運算量，減少解碼時間。 There are increased emphasis in Low-density parity check(LDPC) codes in information theory field. However, one of the main problems is that due to its high decoding complexity in communication system, which leads to much decoding time, it is not acceptable in many applications. We can find that it have good performance in decoding by using message passing algorithm. However, due to its high complexity, which leads to much decoding time, a lot research are studied and discussed how to improve its complexity, which may not decay its performance seriously. Comparing to soft decoding algorithms, hard decoding algorithms are not as outstanding as soft decoding algorithm in the aspect of performance; however there are some advantages such as low complexity, which can be realized in hardware easily and lower decoding time. In this thesis, I mainly discuss that by mixing soft-decision and hard-decision algorithms can we get a balance between performance and decoding time, which can keep its performance in certain level and use less operations to reduce decoding time.