Knowledge discovery in databases is concerned with extracting useful information from databases, and the immune algorithm is a biological theory-based and globally searching algorithm. A specific immune algorithm is designed for discovering a few interesting, high-level prediction rules from databases, rather than discovering classification knowledge as usual in the literatures. Simulations show that this novel algorithm is able to improve the stability of the population, increase the holistic performance and make the rules extracted have higher precision. Knowledge discovery in databases is concerned with the extraction of useful information from databases, and the immune algorithm is a biological theory-based and globally searching algorithm. A specific immune algorithm is designed for discovering a few interesting, high-level prediction rules from databases, rather than discovering classification knowledge as usual in the literatures. Simulations show that this novel algorithm is able to improve the stability of the population, increase the holistic performance and make the rules extracted have higher precision.