The article represents a new approach to increases efficiency and effectiveness of critical Information Security Management Systems (ISMS). The generative unsupervised neural network of Boltzmann Machine use techniques on the input data for analyzing the normal states of the ISMS in order to predict the undesired situations. The international recognized standards define a set of security controls for monitoring, auditing and managing the ISMS. They are distributed in categories and each of them has standardized attributes. Neural units of Boltzmann Machine are based of the security controls infrastructure of ISMS. Integrating neural networks into an ISMS can significantly enhance an organization’s ability to detect, predict, and respond to security abnormal. The NN are largely used in almost all areas in the information and cyber security, but for modeling ISMS behavior as overall complex system, they are a innovative method to reduce costs and time, and to prevent or mitigate of the damages.