<p> Design and optimization of aircraft thermal management systems (TMS) is typically conducted by considering a single system architecture at steady-state conditions, using per?formance metrics such as bleed air flow rate, fuel burn flow rate, or total system mass. However, when trying to increase the overall performance of a legacy system or analyzing new system architectures, it can be difficult to identify how individual component or sub?system changes will propagate throughout the overall TMS. In this thesis, new knowledge and tools are presented that will advance the use of exergy-based design techniques for next generation aircraft thermal management systems (TMS). This is motivated by the fact that exergy destruction is a quantity that can be calculated for any subsystem or component, regardless of energy domain or function. The relationship between exergy destruction min?imization (EDM) and conventional design metrics is investigated and quantified. This is performed through the use of a steady-state analysis and by leveraging a high fidelity model of a complex TMS. It is shown that exergy destruction is not only sensitive to individual component parameters in a manner consistent with conventional performance metrics, but that due to its generalizability, it also captures how changes in one subsystem propagate throughout the overall TMS. Specifically, through a design case study, it is shown that minimizing system-wide exergy destruction rate (without an engine model) yields a similar engine fuel burn rate as when fuel burn is minimized directly, but also results in a signif?icantly lower system mass. Building on these results, a transient design and analysis tool for TMS is developed using a graph theoretic approach. The tool is used on a case study of an air cycle machine (ACM) and on an architecture enumeration case study for a notional TMS. The transient exergy-based analysis is shown to provide insight into how efficiently energy is used at a component level, and captures the differences in thermal performance between architectures. </p>