The artificial intelligence-based systems are getting smarter and more intelligent and find widespread applications on crossing various research fields. The data mining is being increasingly focused currently on the Structural Health Monitoring (SHM) due to big data for the reliability safety engineering and the digital maintenance. The intelligent approaches in the SHM open the effective proactive life cycle management of infrastructure systems. We showed how to realize an intelligent life cycle management applying machine learning which transforms a set of sensor measurements in patterns of the construction safety and health to trace effectively technical functional states of engineering structures. A real dataset of bridge monitoring was used for our application and processed to extract relevant information for forming patterns of functional states of the bridge. Unobvious regularities and dependencies between the patterns can be founded out based on their relative behaviour using fundamental and contemporary machine learning techniques. The behavioural prediction of the patterns in time and the selection of the relevant sensors advance the life cycle management, enable cross-training, intercomparisons of different engineering structures and make it easier supporting optimal human management solutions in the future.