Controversial in its applications and implications, but perfectly rigorous in its mathematics, this book attempts to advance a "general theory of models". The premise is that a phenomenological approach (one which addresses the observed, rather than the underlying, dynamics of a system) is sufficient to achieve a practical understanding of complex systems. The great advantage would be in biological systems, where the consitituent dynamics are of such complexity that only higher-level processes and behaviour can be observed.
This runs counter to the methodology responsible for most major advances in science, and predates the widespread use of computers which allows much greater insight into the complexities, so the work is perhaps out-of-date in this major respect. Moreover, Thom's thinking runs well beyond what experience has shown the mathematics to be capable of, and I would suggest that the book is largely of historical significance. However, it is a provocative and inspiring text, well worth it if you have the the background.
The mathematics on which this book's argument is based is differential topology, which is typically not encountered until the postgraduate level. While the "Mathematical Summary" is a useful review of what you have to know to begin the book, it is of very limited use in actually learning the mathematics, being at best an aide-memoir. For anyone below postgrad, a better start would be Saunder's "Introduction to Catastrophe Theory".