This work presents new intrusion prevention and self-healing system (SH) for critical services network security. The design features of the proposed system are inspired by the human immune system, integrated with pattern recognition nonlinear classification algorithm and machine learning. Firstly, the current intrusions preventions systems, biological innate and adaptive immune systems, autonomic computing and self-healing mechanisms are studied and analyzed. The importance of intrusion prevention system recommends that artificial immune systems (AIS) should incorporate abstraction models from innate, adaptive immune system, pattern recognition, machine learning and self-healing mechanisms to present autonomous IPS system with fast and high accurate detection and prevention performance and survivability for critical services network system. Secondly, specification language, system design, mathematical and computational models for IPS and SH system are established, which are based upon nonlinear classification, prevention predictability trust, analysis, self-adaptation and self-healing algorithms. Finally, the validation of the system carried out by simulation tests