This monograph presents a variety of techniques that can be used for designing robust fault diagnosis schemes for non-linear systems. The introductory part of the book is of tutorial value and serves as a good starting point for newcomers to this field. Subsequently, advanced robust observer structures are presented, along with discussions on parameter estimation-based techniques. Particular attention is drawn to experimental design for fault diagnosis. The book also presents several robust soft computing approaches utilizing evolutionary algorithms and neural networks. All approaches described in this book are illustrated by practical applications.