Since the 1990's regulatory requirements for On-Board-Diagnostics (OBD) have continuously evolved with an increasing application of advanced electronics and control systems that have been adopted for automotive applications. The current and future demands on emissions and performance requirements are pushing the envelope with respect to management of complex control software strategies, hardware components and their interactions. This further challenges the implementation of OBD. In order to build a robust monitor for a complex system which has minimum risk of false detection, a thorough understanding of both system and components is required.In this paper, several methods will be presented that can be utilized to achieve a successful and robust diagnostic system implementation. Implementation begins with a discussion of the major challenges to achieve consistent performance in the base system control. A deep dive approach will provide new methodologies that can be applied to address the necessary diagnostics. This paper will address diagnostic fault handling based on symptoms, statistical methodologies for filtering out noise, and implementation of reference models for diagnostics. It will also discuss redefining the goal of OBD implementation on modern control systems that have multiple I/O as well as coupled and decoupled configurations.