Multi-disciplinary signal systems involve the signal-transforming components in various disciplines. Creative conceptual design of these systems poses great challenge to designers who have usually been taught with the knowledge in very limited disciplines. This paper proposed a conceptual design approach for multi-disciplinary signal systems by intelligently synthesizing signal-transforming knowledge of multi-disciplinary principle solutions. It consisted of a functional knowledge representation model, which can formally represent the signal-transformation knowledge of various multidisciplinary principle solutions, and a sense-selection-action based approach for achieving the synthesis of multi-disciplinary principle solutions. Besudes, a prototype system for achieving computer-aided multi-disciplinary signal transformation conceptual design was developed,  with a knowledge base about various signal-transforming principle solutions. In addition, an example was used to illustrate the proposed approach.