A multi-state system (MSS) is a system in which both the system and its components may exhibit multiple performance levels (or states) varying from perfect operation to complete failure. MSS are usually designed with sufficient redundancies. However, the recovery mechanisms can fail such that an uncovered fault can propagate through the system, and lead to an entire system failure. This phenomenon is known as imperfect coverage (IPC). In this paper, we propose a combinatorial methodology based on multi-state multi-valued decision diagrams (MMDD) and conditional probability for the reliability and sensitivity/importance analysis of MSS subject to IPC. The advantages and application of the MMDD-based method will be illustrated through analysis of an example MSS subject to IPC.