Abstract : The study of cellular automata rules suitable for cryptographic applications is under consideration. On one hand, cellular automata can be used to generate pseudo-random sequences as well as for the design of S-boxes in symmetric cryptography. On the other hand, Boolean functions with good properties like resiliency and non-linearity are usually obtained either by exhaustive search or by the use of genetic algorithms. We propose here to use some recent research in the classification of Boolean functions and to link it with the study of cellular automata rules. As a consequence of our technique, this also provides a mean to get Boolean functions with good cryptographic properties.