High availability of a system can be achieved by performing timely replacement of degraded or failed components. To this end, spare parts are expected to be available and reordered when needed. It is not uncommon that spare parts may deteriorate on the shelf because of their physical characteristics and/or the imperfect storage and transportation conditions. Such phenomena will affect the reliability of spare parts and the availability of the system. In this dissertation, we first focus on a system with single critical operating component and one unit of deteriorating spare part. For such a system, to ensure the system availability and cost efficiency, making a joint decision on component replacement and reordering time is of vital importance. In particular, we study both failure-switching and preventive-switching strategies, where cumulative damage is considered for the spare part switching from its in-stock to operating conditions. To determine the corresponding optimal component replacement and reordering policies, the long-run average costs are minimized under a fixed lead time. It is expected that the work will benefit quite a few industry sectors, such as mining, oil and gas, and defense, where the operation of systems heavily relies on capital-intensive components. To advance the research a step further, we have relaxed the system with only a single operating component to a more complex system with multiple components. In addition, we have eliminated the limitations on the order quantity and inventory capacity. To capture the on-shelf part deterioration, a two-phase deteriorating process is adopted, for which the first phase is from the spare's new arrival to the identification of its degradation, and the second phase is the period thereafter but before the unit fails. Based on the parts' degradation states, we introduce two different replacement strategies for the spare consumption, i.e., the Degraded-First strategy and the New-First strategy. Because of the random nature of component failures and on-shelf deterioration, stochastic cost models for both DF and NF strategies are derived. With the objective of cost reduction through coordinating the inventory and maintenance policies, an enumeration algorithm with stochastic dynamic programming is employed for finding the joint optimal solution over a finite time horizon. Numerical experiments are conducted to study the impacts of these two strategies on the operation costs, and the analysis of key parameters that affect the optimal solutions is also carried out in the numerical study. The joint policies of our interest focus on both replacement and reordering of spare parts, which are more realistic and complex than those policies handling maintenance and spare parts inventory control separately. In particular: When the maintenance planning and inventory control strategy are jointly optimized, we consider the spare parts inventory experiencing on-shelf deterioration, which has not been well studied in the related literature. When dealing with a system carrying only one spare part, the impact of on-shelf deterioration of the spare part on its remaining operational lifetime is explicitly dealt with and described by the Cumulative Exposure (CE) model. For the extended model for a multi-component system, we make an early attempt to adopt a two-phase process to take into account on-shelf degradation of parts. The issues in the degradation-level-based ordering of spare parts in the multi-component system are also discussed. Several integrated cost models are developed in both systems and are used to determine the optimal replacement and reordering decisions with the objective of minimizing the expected long-run cost rate over an infinite/finite horizon.