Recent trends in the development of cyber-physical systems are aimed at their autonomy and self-organization. In this context, the question of monitoring and diagnosing possible evolutionary changes that carry certain risks is very important. This study aims to analyze and determine the approach to monitoring and diagnosing autonomous cyber-physical systems according to their qualitative state. The work of monitoring and diagnostic systems and their interaction with autonomous cyber-physical systems in different scenarios in making decisions about evolutionary transformations by the latter are considered. The ways of further development of monitoring and diagnostic systems based on the assessment of the qualitative state of the studied system are determined.