Proposed evolutionary model provides a powerful means of rule discovery in the field of text categorization. In present paper, the model focuses on the particular problem of preventing spam to enter our e-mail accounts. Spam blockers that have been provided by Internet companies so far are not very effective. An application built through the means of our model learns the approximate difference between spam and non-spam mail and labels incoming new mail efficiently. The rules that led to the membership of training mails to either spam or non-spam category are discovered, having been evolved through the principles of a special evolutionary metaheuristics, genetic chromodynamics. Resulting rules are used to establish the appropriate category for previously unseen mails.