隨著生產技術的提升，產品的複雜性也越來越高，所以產品或系統的可靠與否，是一個重要的課題。而一個產品或系統在操作期間，難以預測的故障往往會造成損失甚至是傷亡，為維持產品或系統的正常操作，適當的維護作業是必須的，因此系統的維護保養是必須考量的問題。 馬可夫過程是一種特別的離散時間隨機過程。首先要將所分析之系統分為有限個狀態，並求得狀態彼此間之轉換機率，然後利用此轉換機率來預測未來某一時間之系統狀態。 本文以動態可靠度為中心，利用馬可夫過程，來探討系統預防維護作業，系統在運轉的壽命期間內，探討系統的參數改變時，動態可靠度與失效率變化情況；可靠度與壽命為系統維護作業的兩個主要因素，彼此之間如此做一權衡，來滿足操作使用的需求。 With improvement of the producing technique, the more complicated the product is. So reliability of a product or a system is the key point. During the operating period of a system or a product, unpredictable accident always causes damage. In order to maintain the routine operation of a product or a system, the proper maintenance is necessary. Therefore, the maintenance of the system must take into count. Markov Process is a special discrete-time stochastic process. First, we have to divide the system we want to analyze into finite states. And find out the transition probability between two states. Finally, this transition probability can be applied to predict the system situation at certain moment in the future. This thesis employs dynamic reliability, and Markov Process to discuss the preventive maintenance of a system. And this thesis discuss the deviation of the dynamic reliability and the failure rate with different parameter of the system during the lifecycle of a system. The reliability and MTBF (Mean Time Between Failure) is two major factor of maintenance. In order to fit the operators’ needs, we have to produce a balance between them.