This paper presents the synchronization and secure communication of chaotic systems through a polynomial fuzzy model. Two novel theorems are proposed for the sum of squares (SOS) stability conditions with a decay rate and an input constraint. The SOS-based framework allows for a number of innovations and improvements over the LMI-based approaches to T-S fuzzy modeling and control. As such, polynomial fuzzy systems are shown to provide more general and effective options than other fuzzy approaches do. Computer simulations illustrate the validity and applicability of this approach.